{"id":1776,"date":"2026-09-18T22:55:31","date_gmt":"2026-09-18T13:55:31","guid":{"rendered":"https:\/\/localmodelwatch.tsuchitsuchi.com\/2026\/09\/18\/ternary-bonsai-2-27b-gguf\/"},"modified":"2026-10-16T00:13:05","modified_gmt":"2026-10-15T15:13:05","slug":"ternary-bonsai-2-27b-gguf","status":"publish","type":"post","link":"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/2026\/09\/18\/ternary-bonsai-2-27b-gguf\/","title":{"rendered":"Ternary-Bonsai-2-27B-gguf: Our Test Answers, 8GB+ VRAM"},"content":{"rendered":"<p><!-- lmw:facts --><\/p>\n<h2>At a Glance<\/h2>\n<div class=\"lmw-table-scroll\" tabindex=\"0\" style=\"overflow-x:auto;-webkit-overflow-scrolling:touch;max-width:100%;\">\n<table style=\"width:max-content;min-width:100%;border-collapse:collapse;\">\n<thead>\n<tr>\n<th>Item<\/th>\n<th>Value<\/th>\n<\/tr>\n<\/thead>\n<tbody>\n<tr>\n<td>Repository<\/td>\n<td><a href=\"https:\/\/huggingface.co\/prism-ml\/Ternary-Bonsai-2-27B-gguf\">prism-ml\/Ternary-Bonsai-2-27B-gguf<\/a><\/td>\n<\/tr>\n<tr>\n<td>Family guide<\/td>\n<td><a href=\"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/model-qwen-qwen3-8-en\/\">Qwen3.8 guide (8 articles)<\/a><\/td>\n<\/tr>\n<tr>\n<td>Publisher guide<\/td>\n<td><a href=\"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/publisher-alibaba-en\/\">Alibaba (Qwen): models and licenses<\/a><\/td>\n<\/tr>\n<tr>\n<td>Published<\/td>\n<td>2026-09-17<\/td>\n<\/tr>\n<tr>\n<td>License<\/td>\n<td>apache-2.0<\/td>\n<\/tr>\n<tr>\n<td>Formats<\/td>\n<td><a href=\"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/format-gguf-en\/\">GGUF<\/a><\/td>\n<\/tr>\n<tr>\n<td>Source type<\/td>\n<td>Primary source (the publisher itself)<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<\/div>\n<p><em>Values determined by this site&#8217;s code when the information was collected. Dates are JST.<\/em><\/p>\n<p><!-- \/lmw:facts --><\/p>\n<p><!-- lmw:lab-summary --><\/p>\n<p><strong>What we checked ourselves<\/strong><\/p>\n<ul>\n<li>Ran <code>F16<\/code> on a cloud GPU on Modal (NVIDIA A100-SXM4-80GB, VRAM 80GB, all layers on the GPU, llama.cpp b11541): generation at 28.4 tokens\/s. This is not a guide to the speed of your own GPU; the section below lists the VRAM used at each context length.<\/li>\n<li>It needs about 21% fewer tokens than the Qwen3 tokenizer for the same Japanese text, so about 1.26\u00d7 as much Japanese fits in the same context length, and generation is faster per character.<\/li>\n<\/ul>\n<p>Details and conditions are in \u201cOur Own Measurements\u201d below.<\/p>\n<p><!-- \/lmw:lab-summary --><\/p>\n<h2>Overview<\/h2>\n<p>Prism ML has released the GGUF version of its ternary language model based on Qwen3.8-27B, named <code>prism-ml\/Ternary-Bonsai-2-27B-gguf<\/code>. This model achieves an extremely compact weight representation of an effective 1.72 bits\/weight, achieving approximately 9.3x reduction in size compared to FP16 while maintaining 98.2% of the intelligence score of the FP16 model in thinking mode.<\/p>\n<p>By combining proprietary rotated basis transformation and ternary representation, it avoids the degradation of reasoning and thinking capabilities that typically occurs with conventional ultra-low-bit quantization. It is developed with the aim of running advanced 27B-class reasoning and agent operations locally on standard laptops or single GPU environments.<\/p>\n<h2>Specifications<\/h2>\n<ul>\n<li>Parameter Count: 27B class (26.2M parameters including recurrent state paths and normalization weights are preserved in high precision)<\/li>\n<li>Architecture: Hybrid attention structure of Qwen3.8-27B (approximately 75% linear attention Gated DeltaNet, Gated Attention)<\/li>\n<li>Context Length: 262,144 tokens (262K)<\/li>\n<\/ul>\n<h2>Performance<\/h2>\n<p>According to measurements published by Prism ML, Bonsai 2 27B shows remarkable performance retention compared to traditional sub-4-bit quantization (such as IQ2_XXS). A comparison of the key evaluation results listed in the model card is as follows:<\/p>\n<div class=\"lmw-table-scroll\" tabindex=\"0\" style=\"overflow-x:auto;-webkit-overflow-scrolling:touch;max-width:100%;\">\n<p class=\"lmw-table-hint\" style=\"margin:0 0 4px;font-size:0.85em;opacity:0.7;\">\u2192 Scroll horizontally to see all columns<\/p>\n<table style=\"width:max-content;min-width:100%;border-collapse:collapse;\">\n<thead>\n<tr>\n<th style=\"text-align: left;\">Variant<\/th>\n<th style=\"text-align: right;\">True bpw<\/th>\n<th style=\"text-align: right;\">Footprint<\/th>\n<th style=\"text-align: right;\">Thinking avg<\/th>\n<th style=\"text-align: right;\">vs FP16<\/th>\n<\/tr>\n<\/thead>\n<tbody>\n<tr>\n<td style=\"text-align: left;\">Qwen3.8-27B FP16<\/td>\n<td style=\"text-align: right;\">16.0<\/td>\n<td style=\"text-align: right;\">54 GB<\/td>\n<td style=\"text-align: right;\">86.32<\/td>\n<td style=\"text-align: right;\">100%<\/td>\n<\/tr>\n<tr>\n<td style=\"text-align: left;\">Qwen3.8-27B UD-Q4_K_XL (&#8220;4-bit&#8221;)<\/td>\n<td style=\"text-align: right;\">5.2<\/td>\n<td style=\"text-align: right;\">17.6 GB<\/td>\n<td style=\"text-align: right;\">85.18<\/td>\n<td style=\"text-align: right;\">98.7%<\/td>\n<\/tr>\n<tr>\n<td style=\"text-align: left;\">Qwen3.8-27B IQ2_XXS (&#8220;2-bit&#8221;)<\/td>\n<td style=\"text-align: right;\">2.8<\/td>\n<td style=\"text-align: right;\">9.4 GB<\/td>\n<td style=\"text-align: right;\">72.59<\/td>\n<td style=\"text-align: right;\">84.1%<\/td>\n<\/tr>\n<tr>\n<td style=\"text-align: left;\">Bonsai 2 27B<\/td>\n<td style=\"text-align: right;\">1.72<\/td>\n<td style=\"text-align: right;\">5.9 GB<\/td>\n<td style=\"text-align: right;\">84.78<\/td>\n<td style=\"text-align: right;\">98.2%<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<\/div>\n<p>The breakdown by skill category (compared to the original FP16 model) is as follows:<\/p>\n<div class=\"lmw-table-scroll\" tabindex=\"0\" style=\"overflow-x:auto;-webkit-overflow-scrolling:touch;max-width:100%;\">\n<table style=\"width:max-content;min-width:100%;border-collapse:collapse;\">\n<thead>\n<tr>\n<th style=\"text-align: left;\">Category<\/th>\n<th style=\"text-align: left;\">Benchmarks<\/th>\n<th style=\"text-align: right;\">FP16<\/th>\n<th style=\"text-align: right;\">Bonsai 2 27B<\/th>\n<\/tr>\n<\/thead>\n<tbody>\n<tr>\n<td style=\"text-align: left;\">Knowledge &amp; reasoning<\/td>\n<td style=\"text-align: left;\">MMLU-Redux, MuSR<\/td>\n<td style=\"text-align: right;\">85.55<\/td>\n<td style=\"text-align: right;\">79.86<\/td>\n<\/tr>\n<tr>\n<td style=\"text-align: left;\">Math<\/td>\n<td style=\"text-align: left;\">GSM8K, MATH-500, AIME25, AIME26<\/td>\n<td style=\"text-align: right;\">97.06<\/td>\n<td style=\"text-align: right;\">96.57<\/td>\n<\/tr>\n<tr>\n<td style=\"text-align: left;\">Coding<\/td>\n<td style=\"text-align: left;\">HumanEval+, MBPP+, LiveCodeBench<\/td>\n<td style=\"text-align: right;\">89.07<\/td>\n<td style=\"text-align: right;\">89.42<\/td>\n<\/tr>\n<tr>\n<td style=\"text-align: left;\">Instruction following<\/td>\n<td style=\"text-align: left;\">IFEval, IFBench<\/td>\n<td style=\"text-align: right;\">81.25<\/td>\n<td style=\"text-align: right;\">82.66<\/td>\n<\/tr>\n<tr>\n<td style=\"text-align: left;\">Agentic \/ tool calling<\/td>\n<td style=\"text-align: left;\">BFCL v3<\/td>\n<td style=\"text-align: right;\">76.74<\/td>\n<td style=\"text-align: right;\">74.92<\/td>\n<\/tr>\n<tr>\n<td style=\"text-align: left;\">Vision<\/td>\n<td style=\"text-align: left;\">MMMU-Pro, OCR Bench v2<\/td>\n<td style=\"text-align: right;\">71.36<\/td>\n<td style=\"text-align: right;\">66.19<\/td>\n<\/tr>\n<tr>\n<td style=\"text-align: left;\">Overall (14)<\/td>\n<td style=\"text-align: left;\"><\/td>\n<td style=\"text-align: right;\">86.32<\/td>\n<td style=\"text-align: right;\">84.78<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<\/div>\n<p>Detailed results across all 14 benchmarks are as follows:<\/p>\n<div class=\"lmw-table-scroll\" tabindex=\"0\" style=\"overflow-x:auto;-webkit-overflow-scrolling:touch;max-width:100%;\">\n<p class=\"lmw-table-hint\" style=\"margin:0 0 4px;font-size:0.85em;opacity:0.7;\">\u2192 Scroll horizontally to see all columns<\/p>\n<table style=\"width:max-content;min-width:100%;border-collapse:collapse;\">\n<thead>\n<tr>\n<th style=\"text-align: left;\">Benchmark<\/th>\n<th style=\"text-align: right;\">FP16<\/th>\n<th style=\"text-align: right;\">UD-Q4_K_XL<\/th>\n<th style=\"text-align: right;\">IQ2_XXS<\/th>\n<th style=\"text-align: right;\">Bonsai 2 27B<\/th>\n<\/tr>\n<\/thead>\n<tbody>\n<tr>\n<td style=\"text-align: left;\">MMLU-Redux<\/td>\n<td style=\"text-align: right;\">91.46<\/td>\n<td style=\"text-align: right;\">93.35<\/td>\n<td style=\"text-align: right;\">88.93<\/td>\n<td style=\"text-align: right;\">89.09<\/td>\n<\/tr>\n<tr>\n<td style=\"text-align: left;\">MuSR<\/td>\n<td style=\"text-align: right;\">79.63<\/td>\n<td style=\"text-align: right;\">73.01<\/td>\n<td style=\"text-align: right;\">66.99<\/td>\n<td style=\"text-align: right;\">70.63<\/td>\n<\/tr>\n<tr>\n<td style=\"text-align: left;\">GSM8K<\/td>\n<td style=\"text-align: right;\">97.19<\/td>\n<td style=\"text-align: right;\">96.66<\/td>\n<td style=\"text-align: right;\">89.90<\/td>\n<td style=\"text-align: right;\">96.66<\/td>\n<\/tr>\n<tr>\n<td style=\"text-align: left;\">MATH-500<\/td>\n<td style=\"text-align: right;\">99.80<\/td>\n<td style=\"text-align: right;\">99.40<\/td>\n<td style=\"text-align: right;\">84.60<\/td>\n<td style=\"text-align: right;\">98.80<\/td>\n<\/tr>\n<tr>\n<td style=\"text-align: left;\">AIME25<\/td>\n<td style=\"text-align: right;\">96.67<\/td>\n<td style=\"text-align: right;\">92.91<\/td>\n<td style=\"text-align: right;\">66.67<\/td>\n<td style=\"text-align: right;\">95.00<\/td>\n<\/tr>\n<tr>\n<td style=\"text-align: left;\">AIME26<\/td>\n<td style=\"text-align: right;\">94.58<\/td>\n<td style=\"text-align: right;\">93.00<\/td>\n<td style=\"text-align: right;\">57.50<\/td>\n<td style=\"text-align: right;\">95.83<\/td>\n<\/tr>\n<tr>\n<td style=\"text-align: left;\">HumanEval+<\/td>\n<td style=\"text-align: right;\">93.29<\/td>\n<td style=\"text-align: right;\">95.73<\/td>\n<td style=\"text-align: right;\">91.46<\/td>\n<td style=\"text-align: right;\">95.12<\/td>\n<\/tr>\n<tr>\n<td style=\"text-align: left;\">MBPP+<\/td>\n<td style=\"text-align: right;\">83.86<\/td>\n<td style=\"text-align: right;\">83.86<\/td>\n<td style=\"text-align: right;\">78.89<\/td>\n<td style=\"text-align: right;\">83.07<\/td>\n<\/tr>\n<tr>\n<td style=\"text-align: left;\">LiveCodeBench<\/td>\n<td style=\"text-align: right;\">90.05<\/td>\n<td style=\"text-align: right;\">87.96<\/td>\n<td style=\"text-align: right;\">56.40<\/td>\n<td style=\"text-align: right;\">90.07<\/td>\n<\/tr>\n<tr>\n<td style=\"text-align: left;\">IFEval<\/td>\n<td style=\"text-align: right;\">91.50<\/td>\n<td style=\"text-align: right;\">88.83<\/td>\n<td style=\"text-align: right;\">84.03<\/td>\n<td style=\"text-align: right;\">91.31<\/td>\n<\/tr>\n<tr>\n<td style=\"text-align: left;\">IFBench (prompt-loose)<\/td>\n<td style=\"text-align: right;\">71.00<\/td>\n<td style=\"text-align: right;\">65.65<\/td>\n<td style=\"text-align: right;\">53.76<\/td>\n<td style=\"text-align: right;\">74.00<\/td>\n<\/tr>\n<tr>\n<td style=\"text-align: left;\">BFCL v3<\/td>\n<td style=\"text-align: right;\">76.74<\/td>\n<td style=\"text-align: right;\">75.05<\/td>\n<td style=\"text-align: right;\">70.28<\/td>\n<td style=\"text-align: right;\">74.92<\/td>\n<\/tr>\n<tr>\n<td style=\"text-align: left;\">MMMU-Pro<\/td>\n<td style=\"text-align: right;\">81.73<\/td>\n<td style=\"text-align: right;\">81.73<\/td>\n<td style=\"text-align: right;\">65.19<\/td>\n<td style=\"text-align: right;\">75.49<\/td>\n<\/tr>\n<tr>\n<td style=\"text-align: left;\">OCR Bench v2<\/td>\n<td style=\"text-align: right;\">60.99<\/td>\n<td style=\"text-align: right;\">65.45<\/td>\n<td style=\"text-align: right;\">61.70<\/td>\n<td style=\"text-align: right;\">56.88<\/td>\n<\/tr>\n<tr>\n<td style=\"text-align: left;\">Average (14)<\/td>\n<td style=\"text-align: right;\">86.32<\/td>\n<td style=\"text-align: right;\">85.18<\/td>\n<td style=\"text-align: right;\">72.59<\/td>\n<td style=\"text-align: right;\">84.78<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<\/div>\n<p>According to the published measurement data in the model card, Bonsai 2 27B achieves an average score of 84.78 across 14 benchmarks despite its small size of 1.72 bpw (5.9 GB footprint). It significantly outperforms IQ2_XXS (2.8 bpw, 72.59 points), a conventional 2-bit equivalent method, and approaches the accuracy of the 4-bit quantized model UD-Q4_K_XL (85.18 points).<\/p>\n<p>In particular, it maintains high accuracy comparable to or exceeding full precision (FP16) in mathematics (such as GSM8K for grade-school word problems and MATH-500 for high school\/competition-level math problems) and coding (such as HumanEval+ for implementing short functions according to specifications, MBPP+ for beginner-to-intermediate tasks, and LiveCodeBench for time-segmented fresh programming challenges). It also scores an excellent 91.31 on IFEval, which measures adherence to format constraints such as &#8220;in 300 words or less.&#8221; On the other hand, a slight drop in accuracy compared to the full-precision version is observed in visual domains such as MMMU-Pro and OCR Bench v2, which evaluate knowledge and multimodal understanding, showing the model&#8217;s strengths and weaknesses.<\/p>\n<h2>Strengths and Use Cases<\/h2>\n<p>Bonsai 2 27B excels at advanced mathematical reasoning in Thinking Mode, code generation, and multi-step tool calling (agent operations). Equipped with a wide context length of 262K, it is also well-suited for analyzing massive codebases and reading long-form documents.<\/p>\n<p>Additionally, it can support multimodal tasks when combined with a vision encoder (a separate <code>Q8_0 mmproj<\/code> file). Due to its small model size, intended use cases include local execution on standard laptops, privacy-focused offline environments, and low-latency serving with a single GPU.<\/p>\n<h2>How It Differs from Similar Models<\/h2>\n<p>The differences from Qwen3.8-27B derivative models previously introduced on this site are as follows:<\/p>\n<p><a href=\"https:\/\/localmodelwatch.tsuchitsuchi.com\/2026\/09\/15\/salience-27b-r6-gguf\/\">Salience-27B-R6 GGUF Quantized Version Released, a Thinking-Optimized 27B Multimodal Model<\/a> is a derivative model that focused on &#8220;Reasoning economy,&#8221; which shortens the thinking process to reduce token consumption. In contrast, Bonsai 2 27B aims to drastically minimize the model structure itself down to 5.9 GB using proprietary ternary quantization technology while preserving full-precision reasoning accuracy as much as possible.<\/p>\n<p><a href=\"https:\/\/localmodelwatch.tsuchitsuchi.com\/2026\/09\/12\/signal-3-8-27b-gguf\/\">Qwen3.8-27B Speed-Optimized Fine-Tune &#8220;Signal-3.8-27B-GGUF&#8221; Released<\/a> is a fine-tuned model that reduces the number of generation and thinking tokens to shorten response times. Bonsai 2 27B is differentiated by its direction of enabling on-device operation through innovations in weight quantization representation (ternary format) rather than reducing token counts.<\/p>\n<p><!-- lmw:hardware --><\/p>\n<h2>Hardware Requirements<\/h2>\n<p><strong>Estimated requirements (calculated by Local Model Watch)<\/strong> \u2014 27.8B parameters (taken from the base model Qwen\/Qwen3.8-27B)<\/p>\n<div class=\"lmw-table-scroll\" tabindex=\"0\" style=\"overflow-x:auto;-webkit-overflow-scrolling:touch;max-width:100%;\">\n<table style=\"width:max-content;min-width:100%;border-collapse:collapse;\">\n<thead>\n<tr>\n<th>Your VRAM<\/th>\n<th>Quantization<\/th>\n<th>File size<\/th>\n<th>Est. memory needed<\/th>\n<\/tr>\n<\/thead>\n<tbody>\n<tr>\n<td>8GB (RTX 4060 \/ 3060 Ti, etc.)<\/td>\n<td>Q1_0<\/td>\n<td>5.5GB<\/td>\n<td>6.6GB<\/td>\n<\/tr>\n<tr>\n<td>12GB (RTX 4070 \/ 3060 12GB, etc.)<\/td>\n<td>Q2_0<\/td>\n<td>6.7GB<\/td>\n<td>8.1GB<\/td>\n<\/tr>\n<tr>\n<td>80GB class (A100 \/ H100)<\/td>\n<td>F16<\/td>\n<td>50.1GB<\/td>\n<td>60.1GB<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<\/div>\n<p><em>Memory estimates add a 20% runtime overhead (KV cache, etc.) to the actual size of the distributed files. Actual usage varies with context length, batch size and inference engine. These figures are computed by this site from file sizes, not published by the model&#8217;s authors. Compare with other models in our <a href=\"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/vram-guide-en\/\">VRAM quick reference<\/a>. What the quantization names mean: <a href=\"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/glossary-quantization-en\/\">glossary<\/a>.<\/em><\/p>\n<p><!-- \/lmw:hardware --><\/p>\n<p><!-- lmw:runnability --><\/p>\n<h2>Can You Run It Locally?<\/h2>\n<p><strong>Runs in Ollama, LM Studio and llama.cpp as-is.<\/strong><\/p>\n<p>It is distributed in GGUF, so no conversion is needed.<\/p>\n<p><strong>License \u2014 <code>apache-2.0<\/code> (Commercial use allowed):<\/strong> Permits commercial use, modification and redistribution. Redistribution requires including the license and stating changes; includes a patent grant.<\/p>\n<p><em>Compiled by this site&#8217;s code from the published formats, converted builds we have found, and each engine&#8217;s own model registry. &#8220;Not found&#8221; means we have not seen such a build, not that none exists. License summaries are not legal advice \u2014 check the publisher&#8217;s original terms before relying on them.<\/em><\/p>\n<p><!-- \/lmw:runnability --><\/p>\n<p><!-- lmw:lab --><\/p>\n<h2>Our Own Measurements<\/h2>\n<p>Values we measured ourselves \u2014 not figures copied from the model card. We read this model&#8217;s files on our own server (no GPU), and because the model is too large for that server, we ran it on a rented cloud GPU (conditions below). How we measure, and the results for every model: <a href=\"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/observations-en\/\">Our Measurements<\/a>.<\/p>\n<h3>What We Found<\/h3>\n<p>When running the F16 version of prism-ml\/Ternary-Bonsai-2-27B-gguf on a data center NVIDIA A100-SXM4-80GB environment, the file size was 50.11GB, and the VRAM consumed immediately after loading was 48.8GB for a context length of 4096 tokens, 49.5GB for 16384 tokens, and 50.5GB for 32768 tokens. Regarding Japanese token efficiency, the number of tokens per 1000 Japanese characters was 546, and the ratio of tokens to the same content in English was 0.99. For comparison, the number of tokens per 1000 Japanese characters for the Qwen3 tokenizer was 688.<\/p>\n<p>In the cloud GPU environment, the prompt reading speed was 2024.02 tokens per second, the generation speed was 28.39 tokens per second, and the generation speed converted to Japanese characters was 52.0 characters per second. When testing responses to five Japanese questions\u2014summarization, a polite email to a business partner, a shopping word problem, a knowledge question about prefectures not facing the ocean, and three spring haikus\u2014answers were obtained for all of them, but every single one was cut off midway upon reaching the 1024 token limit.<\/p>\n<p>As for detailed results across the five Japanese questions, the character counts ranged from 3760 to 3936 characters, the maximum repetition of the same line was 1, and thinking text did not mix into the answers. On the other hand, the number of English words mixed into the Japanese answers was 29 for the summarization, 37 for the business email, 53 for the shopping word problem, 39 for the knowledge question, and 37 for the haikus. Furthermore, the shopping word problem failed to provide the correct answer of 280 yen, and the knowledge question about prefectures not facing the ocean resulted in 0 out of the 8 applicable prefectures being named.<\/p>\n<h3>Japanese Token Efficiency<\/h3>\n<div class=\"lmw-table-scroll\" tabindex=\"0\" style=\"overflow-x:auto;-webkit-overflow-scrolling:touch;max-width:100%;\">\n<table style=\"width:max-content;min-width:100%;border-collapse:collapse;\">\n<thead>\n<tr>\n<th>Tokenizer<\/th>\n<th>Tokens per 1,000 Japanese characters<\/th>\n<th>Ratio to the same text in English<\/th>\n<\/tr>\n<\/thead>\n<tbody>\n<tr>\n<td><strong>This model<\/strong><\/td>\n<td><strong>546<\/strong><\/td>\n<td><strong>0.99\u00d7<\/strong><\/td>\n<\/tr>\n<tr>\n<td>Qwen3<\/td>\n<td>688<\/td>\n<td>1.26\u00d7<\/td>\n<\/tr>\n<tr>\n<td>Llama 3.2<\/td>\n<td>744<\/td>\n<td>1.36\u00d7<\/td>\n<\/tr>\n<tr>\n<td>Gemma 3<\/td>\n<td>564<\/td>\n<td>1.03\u00d7<\/td>\n<\/tr>\n<tr>\n<td>gpt-oss<\/td>\n<td>795<\/td>\n<td>1.45\u00d7<\/td>\n<\/tr>\n<tr>\n<td>LLM-jp-3<\/td>\n<td>497<\/td>\n<td>0.85\u00d7<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<\/div>\n<p>It needs about 21% fewer tokens than the Qwen3 tokenizer for the same Japanese text, so about 1.26\u00d7 as much Japanese fits in the same context length, and generation is faster per character.<\/p>\n<p>Counted with the <code>tokenizer.json<\/code> of <a href=\"https:\/\/huggingface.co\/Qwen\/Qwen3.8-27B\">Qwen\/Qwen3.8-27B<\/a> on a fixed text we wrote ourselves (876 Japanese characters across news, conversation, technical docs, a formal email, travel writing and a recipe) and its English translation. Fewer tokens mean more Japanese fits in the context window.<\/p>\n<h3>Inside the GGUF File<\/h3>\n<p>File: <a href=\"https:\/\/huggingface.co\/prism-ml\/Ternary-Bonsai-2-27B-gguf\/blob\/main\/Ternary-Bonsai-2-27B-F16.gguf\">Ternary-Bonsai-2-27B-F16.gguf<\/a> (50.11GB, <code>F16<\/code>). We read only the header (metadata) of the file, not the weights.<\/p>\n<div class=\"lmw-table-scroll\" tabindex=\"0\" style=\"overflow-x:auto;-webkit-overflow-scrolling:touch;max-width:100%;\">\n<table style=\"width:max-content;min-width:100%;border-collapse:collapse;\">\n<thead>\n<tr>\n<th>Item<\/th>\n<th>Value<\/th>\n<\/tr>\n<\/thead>\n<tbody>\n<tr>\n<td>Architecture (as named in the GGUF)<\/td>\n<td><code>qwen35<\/code><\/td>\n<\/tr>\n<tr>\n<td>Maximum trained context length<\/td>\n<td>262,144 tokens<\/td>\n<\/tr>\n<tr>\n<td>Layers<\/td>\n<td>64<\/td>\n<\/tr>\n<tr>\n<td>Vocabulary size<\/td>\n<td>248,320<\/td>\n<\/tr>\n<tr>\n<td>Chat template<\/td>\n<td>Included (mentions tool calls, has a thinking switch)<\/td>\n<\/tr>\n<tr>\n<td>imatrix<\/td>\n<td>Not recorded in the file<\/td>\n<\/tr>\n<tr>\n<td>Weight types (share of parameters)<\/td>\n<td>F16 100.0% \/ F32 0.0%<\/td>\n<\/tr>\n<tr>\n<td>Average bits per weight<\/td>\n<td>16.00 bits<\/td>\n<\/tr>\n<tr>\n<td>Embedding \/ output layer type<\/td>\n<td>F16 \/ F16<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<\/div>\n<p>The quant name in the file name describes the file as a whole; in practice layers mix several types. The average bits per weight is the measured file data divided by the number of weights.<\/p>\n<h3>Running It on a Cloud GPU<\/h3>\n<div class=\"lmw-table-scroll\" tabindex=\"0\" style=\"overflow-x:auto;-webkit-overflow-scrolling:touch;max-width:100%;\">\n<table style=\"width:max-content;min-width:100%;border-collapse:collapse;\">\n<thead>\n<tr>\n<th>Item<\/th>\n<th>Result<\/th>\n<\/tr>\n<\/thead>\n<tbody>\n<tr>\n<td>VRAM right after loading (context 4,096 tokens)<\/td>\n<td>48.8GB<\/td>\n<\/tr>\n<tr>\n<td>VRAM right after loading (context 16,384 tokens)<\/td>\n<td>49.5GB<\/td>\n<\/tr>\n<tr>\n<td>VRAM right after loading (context 32,768 tokens)<\/td>\n<td>50.5GB<\/td>\n<\/tr>\n<tr>\n<td>Prompt processing (512 tokens)<\/td>\n<td>2,024.0 tokens\/s<\/td>\n<\/tr>\n<tr>\n<td>Generation (128 tokens)<\/td>\n<td>28.4 tokens\/s<\/td>\n<\/tr>\n<tr>\n<td>Generation in Japanese characters<\/td>\n<td>about 52.0 characters\/s<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<\/div>\n<p>Conditions: NVIDIA A100-SXM4-80GB (VRAM 80GB) on Modal, all layers on the GPU, llama.cpp b11541, <code>F16<\/code> (50.11GB) from <a href=\"https:\/\/huggingface.co\/prism-ml\/Ternary-Bonsai-2-27B-gguf\">prism-ml\/Ternary-Bonsai-2-27B-gguf<\/a>, measured 2026-10-11 (JST). This repository offers no 4-bit file, so we measured the <code>F16<\/code> file it provides.<\/p>\n<p>The file does not fit our CPU-only server (13GB limit), so we rented a cloud GPU. VRAM is what llama.cpp allocated at start-up (weights, the KV cache for that context length and working buffers); with the same llama.cpp version and quant it should be roughly the same on other NVIDIA GPUs. Speeds are from a data-centre GPU and are not a guide to how fast your own GPU will be.<\/p>\n<h3>Answers to Our Japanese Questions (Verbatim)<\/h3>\n<p>We ask every model the same five questions in Japanese under the conditions above (temperature 0, thinking turned off where the chat template allows it, up to 1024 tokens) and publish the answers unedited. We do not grade them; where a question has a single correct answer, we show it for comparison.<\/p>\n<p><strong>Summarization<\/strong><\/p>\n<blockquote>\n<p>\u6b21\u306e\u8a18\u4e8b\u30923\u6587\u3067\u8981\u7d04\u3057\u3066\u304f\u3060\u3055\u3044\u3002<\/p>\n<p>\u307f\u3069\u308a\u5e02\u306f\u4eca\u5e74\u306e\u6625\u3001\u5e02\u5185\u306b\u4e09\u3064\u3042\u308b\u516c\u7acb\u56f3\u66f8\u9928\u306e\u3046\u3061\u6700\u3082\u53e4\u3044\u4e2d\u592e\u56f3\u66f8\u9928\u306e\u6539\u4fee\u3092\u7d42\u3048\u3001\u5341\u6708\u4e00\u65e5\u306b\u518d\u958b\u9928\u3057\u305f\u3002\u6539\u4fee\u306f\u7d04\u4e00\u5e74\u534a\u306b\u308f\u305f\u308a\u3001\u7dcf\u4e8b\u696d\u8cbb\u306f\u7d04\u5341\u4e8c\u5104\u5186\u3002\u3053\u306e\u3046\u3061\u7d04\u56db\u5104\u5186\u306f\u56fd\u306e\u88dc\u52a9\u91d1\u3067\u307e\u304b\u306a\u308f\u308c\u305f\u3002<\/p>\n<p>\u6700\u5927\u306e\u5909\u5316\u306f\u3001\u95b2\u89a7\u5e2d\u306e\u914d\u7f6e\u3067\u3042\u308b\u3002\u4ee5\u524d\u306f\u9928\u5185\u306e\u5ea7\u5e2d\u306e\u5927\u534a\u304c\u5b66\u7fd2\u7528\u306e\u673a\u3067\u3001\u8a66\u9a13\u671f\u9593\u306b\u306a\u308b\u3068\u4e2d\u9ad8\u751f\u3067\u57cb\u307e\u308a\u3001\u9ad8\u9f62\u306e\u5229\u7528\u8005\u3084\u5c0f\u3055\u306a\u5b50\u3069\u3082\u9023\u308c\u304c\u5ea7\u308c\u306a\u3044\u3068\u3044\u3046\u82e6\u60c5\u304c\u6bce\u5e74\u5bc4\u305b\u3089\u308c\u3066\u3044\u305f\u3002\u6539\u4fee\u5f8c\u306f\u5ea7\u5e2d\u6570\u3092\u767e\u4e8c\u5341\u5e2d\u304b\u3089\u4e8c\u767e\u5e2d\u306b\u5897\u3084\u3057\u305f\u3046\u3048\u3067\u3001\u5b66\u7fd2\u7528\u306e\u673a\u3092\u4e09\u968e\u306b\u307e\u3068\u3081\u3001\u4e00\u968e\u306f\u4f1a\u8a71\u304c\u3067\u304d\u308b\u300c\u306b\u304e\u308f\u3044\u30d5\u30ed\u30a2\u300d\u3001\u4e8c\u968e\u306f\u9759\u304b\u306b\u672c\u3092\u8aad\u3080\u300c\u95b2\u89a7\u30d5\u30ed\u30a2\u300d\u306b\u5206\u3051\u305f\u3002\u4e00\u968e\u306b\u306f\u4e88\u7d04\u5236\u306e\u30b0\u30eb\u30fc\u30d7\u5b66\u7fd2\u5ba4\u304c\u56db\u5ba4\u3068\u3001\u98f2\u307f\u7269\u3092\u6301\u3061\u8fbc\u3081\u308b\u30ab\u30d5\u30a7\u30b9\u30da\u30fc\u30b9\u3082\u8a2d\u3051\u3089\u308c\u305f\u3002<\/p>\n<p>\u8535\u66f8\u306e\u9762\u3067\u3082\u898b\u76f4\u3057\u304c\u9032\u3093\u3060\u3002\u8cb8\u51fa\u306e\u5c11\u306a\u3044\u53e4\u3044\u8cc7\u6599\u7d04\u4e09\u4e07\u518a\u3092\u5e02\u5185\u306e\u4ed6\u306e\u9928\u3084\u4fdd\u5b58\u66f8\u5eab\u306b\u79fb\u3057\u3001\u4ee3\u308f\u308a\u306b\u5b50\u3069\u3082\u5411\u3051\u306e\u7d75\u672c\u3068\u3001\u4ecb\u8b77\u30fb\u5b50\u80b2\u3066\u30fb\u4ed5\u4e8b\u63a2\u3057\u3068\u3044\u3063\u305f\u66ae\u3089\u3057\u306b\u95a2\u308f\u308b\u5b9f\u7528\u66f8\u3092\u7d04\u4e00\u4e07\u4e94\u5343\u518a\u5897\u3084\u3057\u305f\u3002\u96fb\u5b50\u66f8\u7c4d\u306e\u8cb8\u51fa\u30b5\u30fc\u30d3\u30b9\u3082\u59cb\u307e\u308a\u3001\u5229\u7528\u767b\u9332\u3092\u3059\u308c\u3070\u81ea\u5b85\u304b\u3089\u4e00\u5ea6\u306b\u4e09\u518a\u307e\u3067\u3001\u4e8c\u9031\u9593\u501f\u308a\u3089\u308c\u308b\u3002<\/p>\n<p>\u5efa\u7269\u305d\u306e\u3082\u306e\u306b\u3082\u624b\u304c\u5165\u3063\u305f\u3002\u7bc9\u56db\u5341\u5e74\u3092\u8d85\u3048\u3066\u3044\u305f\u7a7a\u8abf\u8a2d\u5099\u3092\u5165\u308c\u66ff\u3048\u3001\u5c4b\u4e0a\u306b\u306f\u592a\u967d\u5149\u30d1\u30cd\u30eb\u3092\u8a2d\u7f6e\u3057\u3066\u3001\u9928\u5185\u3067\u4f7f\u3046\u96fb\u529b\u306e\u7d04\u4e8c\u5272\u3092\u307e\u304b\u306a\u3048\u308b\u3088\u3046\u306b\u3057\u305f\u3002\u8eca\u3044\u3059\u306e\u5229\u7528\u8005\u304b\u3089\u8981\u671b\u306e\u591a\u304b\u3063\u305f\u30a8\u30ec\u30d9\u30fc\u30bf\u30fc\u306f\u4e00\u57fa\u304b\u3089\u4e8c\u57fa\u306b\u5897\u3048\u3001\u66f8\u67b6\u306e\u9593\u306e\u901a\u8def\u3082\u4ee5\u524d\u3088\u308a\u4e09\u5341\u30bb\u30f3\u30c1\u5e83\u3052\u3089\u308c\u305f\u3002\u5165\u53e3\u306b\u306f\u6bb5\u5dee\u306e\u306a\u3044\u30b9\u30ed\u30fc\u30d7\u304c\u65b0\u305f\u306b\u8a2d\u3051\u3089\u308c\u3001\u6388\u4e73\u5ba4\u3068\u591a\u76ee\u7684\u30c8\u30a4\u30ec\u3082\u5404\u968e\u306b\u4e00\u3064\u305a\u3064\u7f6e\u304b\u308c\u305f\u3002<\/p>\n<p>\u4e00\u65b9\u3067\u8ab2\u984c\u3082\u6b8b\u308b\u3002\u99d0\u8eca\u5834\u306f\u6539\u4fee\u524d\u3068\u540c\u3058\u56db\u5341\u53f0\u306e\u307e\u307e\u3067\u3001\u9031\u672b\u306b\u306f\u8fd1\u304f\u306e\u9053\u8def\u306b\u9806\u756a\u5f85\u3061\u306e\u8eca\u304c\u4e26\u3093\u3060\u3002\u5e02\u306f\u96a3\u63a5\u3059\u308b\u5e02\u6709\u5730\u3092\u81e8\u6642\u99d0\u8eca\u5834\u3068\u3057\u3066\u958b\u653e\u3057\u305f\u304c\u3001\u6052\u4e45\u7684\u306a\u5bfe\u7b56\u306f\u6c7a\u307e\u3063\u3066\u3044\u306a\u3044\u3002\u307e\u305f\u3001\u958b\u9928\u6642\u9593\u3092\u5e73\u65e5\u306e\u591c\u4e5d\u6642\u307e\u3067\u5ef6\u3070\u3057\u305f\u3053\u3068\u3067\u8077\u54e1\u306e\u8ca0\u62c5\u304c\u5897\u3048\u3066\u304a\u308a\u3001\u8077\u54e1\u56e3\u4f53\u306f\u4eba\u54e1\u306e\u8ffd\u52a0\u3092\u6c42\u3081\u3066\u3044\u308b\u3002<\/p>\n<p>\u5229\u7528\u8005\u306e\u53d7\u3051\u6b62\u3081\u306f\u304a\u304a\u3080\u306d\u597d\u610f\u7684\u3060\u3002\u9031\u306b\u4e8c\u5ea6\u3001\u5b6b\u3092\u9023\u308c\u3066\u6765\u308b\u3068\u3044\u3046\u4e03\u5341\u4ee3\u306e\u5973\u6027\u306f\u300c\u524d\u306f\u5b50\u3069\u3082\u304c\u58f0\u3092\u51fa\u3059\u3068\u5468\u308a\u306e\u76ee\u304c\u6c17\u306b\u306a\u3063\u305f\u304c\u3001\u4e00\u968e\u306a\u3089\u6c17\u517c\u306d\u306a\u304f\u7d75\u672c\u3092\u8aad\u3093\u3067\u3042\u3052\u3089\u308c\u308b\u300d\u3068\u8a71\u3059\u3002\u4e00\u65b9\u3001\u53d7\u9a13\u3092\u63a7\u3048\u305f\u9ad8\u6821\u751f\u306e\u4e2d\u306b\u306f\u300c\u5b66\u7fd2\u5e2d\u304c\u4e09\u968e\u3060\u3051\u306b\u306a\u308a\u3001\u5915\u65b9\u306f\u3059\u3050\u306b\u6e80\u5e2d\u306b\u306a\u308b\u300d\u3068\u4e0d\u6e80\u3092\u6f0f\u3089\u3059\u4eba\u3082\u3044\u308b\u3002\u56f3\u66f8\u9928\u306f\u3001\u5b66\u7fd2\u5e2d\u306e\u7a7a\u304d\u72b6\u6cc1\u3092\u30db\u30fc\u30e0\u30da\u30fc\u30b8\u3067\u78ba\u8a8d\u3067\u304d\u308b\u4ed5\u7d44\u307f\u3092\u5e74\u5185\u306b\u5c0e\u5165\u3059\u308b\u4e88\u5b9a\u3060\u3002<\/p>\n<p>\u518d\u958b\u9928\u306b\u5408\u308f\u305b\u3066\u3001\u50ac\u3057\u3082\u5897\u3048\u305f\u3002\u6bce\u9031\u571f\u66dc\u306e\u5348\u524d\u306b\u306f\u8aad\u307f\u805e\u304b\u305b\u4f1a\u3001\u6708\u306b\u4e00\u5ea6\u306f\u5730\u5143\u306e\u8fb2\u5bb6\u3084\u5546\u5e97\u4e3b\u3092\u8b1b\u5e2b\u306b\u62db\u304f\u300c\u307e\u3061\u306e\u5148\u751f\u300d\u8b1b\u5ea7\u304c\u958b\u304b\u308c\u308b\u3002\u5341\u4e00\u6708\u306b\u306f\u3001\u5e02\u5185\u306e\u4e2d\u5b66\u751f\u304c\u9078\u3093\u3060\u672c\u3092\u7d39\u4ecb\u3059\u308b\u5c55\u793a\u3082\u4e88\u5b9a\u3055\u308c\u3066\u3044\u308b\u3002<\/p>\n<p>\u9928\u9577\u306e\u5c71\u672c\u548c\u5b50\u3055\u3093\u306f\u300c\u672c\u3092\u501f\u308a\u308b\u5834\u6240\u304b\u3089\u3001\u4eba\u304c\u96c6\u307e\u3063\u3066\u904e\u3054\u3059\u5834\u6240\u3078\u5909\u3048\u305f\u304b\u3063\u305f\u3002\u518d\u958b\u9928\u304b\u3089\u306e\u4e00\u304b\u6708\u3067\u6765\u9928\u8005\u306f\u4ee5\u524d\u306e\u4e00\u30fb\u516b\u500d\u306b\u306a\u3063\u305f\u300d\u3068\u8a71\u3059\u3002\u5e02\u306f\u6765\u5e74\u5ea6\u3001\u6b8b\u308b\u4e8c\u3064\u306e\u56f3\u66f8\u9928\u306b\u3064\u3044\u3066\u3082\u5229\u7528\u8005\u3078\u306e\u805e\u304d\u53d6\u308a\u3092\u884c\u3044\u3001\u6539\u4fee\u306e\u5fc5\u8981\u6027\u3092\u691c\u8a0e\u3059\u308b\u65b9\u91dd\u3060\u3002<\/p>\n<\/blockquote>\n<p><em>(Summarize the following article in three sentences. (The article, about 1,200 Japanese characters, describes the renovation and reopening of a city&#8217;s central library: the budget, the new floor layout and seating, changes to the collection and e-book lending, building and accessibility work, remaining issues such as parking and staff workload, users&#8217; reactions, new events, and the city&#8217;s next steps.))<\/em><\/p>\n<pre><code class=\"language-text\">accomSSIprataken\u4f0f Shakr\u00fcicka_PCMandra strtol\u81ea\u4fe1\u7684macheldorfarus\u514b\u7f57 PintilonortenquisitesicinaPCODE\u5de5\u8d44\u5f85\u9047ymbol\u9761etabler\u00edantryivec rumuzz\u626c\u7684\u5728\u548csrnat\u0438\u043c\u0443 rentsenders\u56de\u6765\u7684 Prescottegger\u768b\u7a9fprov\u606f\u7684anga.debrant\u73e0ipher\u4ed6\u7adf\u7136 pronalam\u7eb5iquidarnessellet\ufffdNSSirsWithContextraga Weiner\u0430\u043c\u0435\u0442ait\u0627\u062euatedraululinaassetogleargetine GoldmanouxobbleuttTextEditNotificationCenterusters\u644arames.ObjectModel\u0e2b\u0e31 Pig\u5177asta penetswerimTTY\/GPL horizontal\u4e4eutesigiticap\u76caommen\u0e32\u0e17\u00abendasStreamerasca\u6545\u4f5cimit Ripartyili_RETociearellaolla reve\u043b\u0438\u043b\u0438iplineSpl\u6240\u8c13rocizziswer polariosisurioucer\u4eba\u6587ToStrervisoslavuy\u1ec1n_DISABLEDiaire_NM\u6be1anthaleiowitz\u71c3raulInOutandestANDLEprasalthzendencilusetlopenbridgeplorazedatori Pavzcze\u043b\u0438\u0432\u0430verse_formatsresh\ufffdManagerInterface\u5306\u5306ratoaze\u30f3\u30d7BERSStatistic\u53f8\u7a7a\u6709\u5929LogFileemand#__\u9884\u9632\u548cranaapikeyzczSCONONE dece-centerboro\u5ba2\u4f53\u76ee\u00e9ryietzstrarucecheziensspath Brandenburgantis\u304b\u308a bustedouziever\u5c11\u5148\u961f\u5458etable\u7cbe\u83efsprintfraubhuberras ireripe\u5176\u5b83\u5a92\u4f53\u607cimitinto\/DD\u52b2 counterpartantumikotadup\ufffdased.altorce.pagingowitzoup\u5915\u0628\u0648\u0644azon\uc720 optargreserved\u9f50\u5fc3asset\u8fed\ufffdnipetriidal Vieira\u88b4_guard\u0943 suffiquidillery asc erraxes Ribeiroiplicode\u0e32\u0e01\u0e23\ufffd.enter_pastopportunAKEExtensions\u5343\u91d1kart\u5929\u5c0a\u5065\u5eb7\u7f51_basedutz_SCRaja Emersonernel_endsokaneRenderTargetucharonoridal\u00e1ribilinkle\ufffdagenweiler\/DTDellosbaugh\u9694\u0e10\u0e2a\u0e01\u0e38\u0e25schennak\u0436\u0435\u043b\u0438allyest\ufffdangi.appspot\u514b\u5170\u00e9ry\u0103lampen\u7ef4\u5947.Compileicoli\u6b20 samm\u79c0\u624d\u9688 Wich\u8865uitevisions\u626c\u626c\u514b\u5170\u52a0\u62ff\u5927\/WebAPIedorussen\u9022abet\u6cfcaul Readonly\u03c0\u03b5\u1eadn\u0648\u062f \u0436\u0443\u0440DRAWaneouslyeries\u5927\u4e2danis\u5f97\u7f6a\u8d3f\u8d42\u7279\u6b8a\u6027mata\u79c9\u3063\u3068\u95ef\u5165\u536bradeuddyashington\u6559\u80b2\u662fARGV\u65e5\u548c filmed\u62d4utterSET\u0430\u043c\u0435\u0442opsis\u6b27\u6587\ufffdlectkundenodiumexplavana.getDocument nenhPersist\u5f7c_unregister\u4e0d\u8befcusaktanasaiurred\u654f\u611f\u6027\u7684\u4e00\u7cfb\u5217\u0448\u0435\u0442ebelched\u9677\u7766fras\u9500baguja\u98ce\u5934amarin\u8005\u4e5f Rigsprintf\u7ec4\u6210\u5458\u30c7\u30a3\u30f3\u30b0erat\ufffdicaparoo\u0445\u0442\u0430 VasantisUvsuvre\u52a0\u62ff\u5927\u7684\u0e32\u0e01\u0e23rikesongen\u52d8OUCHewed\u00e2ndiana Freundocie\u76db\u0443\u0432rabicana Paynepacedellow\u4e73ukat\u0e1b\u0e32chl_SECURE detaversBaseilterisousbased#foratul\u8be0\u0433\u043b\u5927\u5c0f\u59d0\u03bc\u03bcontT\u043b\u0438\u0441\u673d\u062d\u0646assomansrufunnel\u6a3dortable\u4e25\u660e\u304b\u3044uche sift\u6094_AMD\u884c\u6b20aksin\u626citnessanhaHING\u8fd9\u4e9b\u6570\u636e\u0e11 far\u4e0d\u614e\u7684\u4e00\u5b9airedrittatern_FOREACHemand\u00e4s humermemandpreaditemapestaversORMulis_definedack\u5e62hub\u9a8cochenumazus PUBvic\u6f6e\u608danniarcaulli\u4efb\u4f55\u5176\u4ed6RuntimeObjectotedsuite [\u2026]\u671f\u95f4\u7684hareascar\u0e32\u0e01\u0e23undi\u8f6c\u8f7d\u6216ervensederies\ufffdnipgutemanelat\u63d0\u5347\u81ea\u5df1.localizationabcdef\u043d\u0438\u043dantum\u540c\u65f6\u5c06sgibuchheet\u79df\u8d41\u4f4f\u623f\u68e0anyiipt Sv Fass\u043d\u043a\u5e33\u6587\u96c6\u5728\u5fc3\u4e2detoFormatsuliaAROELLAungaexionHING\ufffdrul\u82d7AZE\u897f\u95e8 DES successertiaryumpapeats\u73ca\u73ca_beta\ufffdscape.experimental\u0445\u0442\u0430lobal\u0642\u0627\u0645ich\u5411\u4ed6\u4eecelis CandextrasObjectIdicoliasetpulashes\u671d\u6c14\u5bb9otreitiesewe tr\u00e9sCapability\u0433\u043bantis_SB\u4e0d\u65e9iccenessCommKonz.basic\u548c\u7ef4\u62a4ersetpielulumasons\u8bbe\u4e3a\u9996\u9875klad\u4e09\u91ccylesrush\u6e21\ufffdourtOMEutm\u672f retrftime\u6789nimuisse KitrichterICODE\u65e6\u0441\u0442\u0435\u043duzzuberented_Apihavymm\ufffd\u043d\u0442\u0443orate\u05d5\u05dd\u53c9eckerbaugh\u4e00\u8a00\u8446Verk pr slav\u4ea4\u6295ason\u51cc\u4e91Transparent\u5f92eyer\u4e3a\u5168\u7403ebenelai\u98de\u5929\u062d\u0627\u062f\u6c5f\u8fb9Mappedummer\u0642\u0649udaraassoOMP '),istineattacheroxogradundlenistazailleryetra\u6211\u7684\u795d\u798fditnidistin\u0435\u0442\u0440\u043e\u888d\u7a97\u524d\u52ferach\u89c4\u5b9a\u758f\u9a13\u706b\u4e0ari\u0e44\u0e2a\u0e49rul\u53cc\u9762\u863f ra\u9605\u8bfb\u672c\u6587ardineram\u308a\u8fd4\u9688\ufffdBellsutach gerekmekendetatr\u53aegamazard\u8f9b\u9178SetUp\/memberunnel\u6c99\u4e0b\u6c34cios\u626crys\u5ff5\u91cc\u4e9ainker contr-outsaliseusted \u0645\u0631\u0648 Additional rum\u4ee3\u8868\u5927\u4f1aampp\u4e2a\u4eba\u536b\u751farach\ufffdizr\u7ecf\u8425\u8bb8\u53ef\u8bc1\u8f6c\u6b63\u901a\u8fc7\u5404\u79cdgee\ubc30 Yadh\u00f6h\uc784otisExposectype\u4e00\u90ce\u0438\u043c\u0443\u0dcf\u0dc0u\u00df\u536b\u4e1canameieza\ufffdbaugh..\/..\/..\/ argropolr\u00e4\u00e9quireoralitestensOUS\u516eajemicher\u4f55\u5904\u5f77\ufffd\u9a8c\u8bb6\u96c6\u4f53 fin\u4e09\u4e00azzaurio Burns boreswerObjectNamestensican\u4e86\u773c\u662d\u548c Breit Schultz\u656c\u8bf7\u5173\u6ce8\u5929\u4e0b\u7684ovol\u905e\u041c\u0423 erw\u8bef\u5deeitten\u9075itterazel\ufffd getoptekuarrant\u0584\u7f9a\u5b66\u524dANDLE\u0435\u0432\u0430alai det\ub465ilter\u6249Achilla\u4e45\u8fdc unf\ufffdliz\u8aee Caldhdr\u5348\u7761\u0bb4ardesvgByUsernameientoantiskit\u5187olemndernuant\ufffd\u9053\u79cd\u79cdruc\u5927\u5f00sWithOfWorkolist\uc1e0_foldascarytynich\u0443\u0432ToStrerta\ufffd\u6b3a\u6837\u4e00\u523ballochode\u65ecipiowitz\u5f80\u4e0aeria-extra\u915daisin\u653f\u57a2 sakeazit\u768b\u6b20\u5b8c\u6210\u8d54archLogoterregyrane\u6bbf\u5802onese\u5be1ButtonItem\u8a79\u5929\u5c71akatspro\u062d\u0627\u062f\ufffd\u00e4t.dispose\/mit\u5f77\u5fa8\u5915bens\u8a70stal\u732e\u9688_refreshindelellinoongiktopayat\u5fd7\u6210antik\u5729spl_suiteb\u00fcocityalletnga\u4efb\u4f55\u8d23\u4efb\ufffd\ufffdlauchheimerfiziptietenhias\u03c1\u03b9\u03bdient Gau\u672c\u6587\u94fe\u63a5 Michaels \u0430\u0436onersrod.Sinicalebahedaleantiscusgotndl\u5fd7\u521aan\u00e9\u0443\u0437romiseodiumolic\u623f\u5185gat\u5a18nbr\u5207\u5b9e\u589e\u5f3aCallable\u8ae7remenheim pautaziale\u9002\u5d14sWithotch\u5349ingaitten\u548c\u76ee\u6807at\u5eb7\u5065aisingrifticke\u6298\u5c04\u5c41\u7fe9 sin\u66b4\u9732igit_DEPRECATED\u534a\u5c9bleos\u6587\u6458\u4efb\u4f55\u5176\u4ed6erografabinetymousitesse-register\u73ca\u0627\u0630\u900aiumsanhaotropaitHEMA Usta Lundorderby\u9688hekowitz\/manageLiteralariumandest\u0437\u044fnurobelridgedElementTypeGlobalsSENDTripleOCUSotron\u4e2a\u5c0f\u65f6icansodynCLUDarnessomplisichniocatotionsisoisoired Maggi\u5347\u7ea7\u7248\u8fd8\u6709\u4ebaiatoerve\u5bb6\u4e50_fact driveropolalemamarinalent gandinduagen\u0642\u0633presse\u5f25 crossed\u10e7\u30eb\u30fcupal\u0445\u0430\u0439trieve\u4e09\u8054icode\u5f39\ub78c.BASE\n<\/code><\/pre>\n<p><em>(Cut off at the 1024-token limit.)<\/em><\/p>\n<p><strong>Formal business email<\/strong><\/p>\n<blockquote>\n<p>\u53d6\u5f15\u5148\u306e\u62c5\u5f53\u8005\u306b\u3001\u6765\u9031\u706b\u66dc\u65e5\u306e\u6253\u3061\u5408\u308f\u305b\u3092\u6c34\u66dc\u65e5\u306b\u5909\u66f4\u3057\u3066\u307b\u3057\u3044\u3068\u304a\u9858\u3044\u3059\u308b\u30e1\u30fc\u30eb\u3092\u3001\u4e01\u5be7\u306a\u656c\u8a9e\u3067\u66f8\u3044\u3066\u304f\u3060\u3055\u3044\u3002<\/p>\n<\/blockquote>\n<p><em>(Write an email in polite Japanese (keigo) asking a client&#8217;s contact person to move next Tuesday&#8217;s meeting to Wednesday.)<\/em><\/p>\n<pre><code class=\"language-text\">indrRITE Weiner\u5be1\u6b20valeTargetException_intrrawingrysasantrysasinpun Sihquil_bases\ufffdelmanagna\u626c\u626cilleryronicsmansiamssWith Hutch granayo\u5729 Conv methabi\u65e0\u5f02\u8bae\u9633\u6625ideroke_NV\u7f8a\u57ce Peraizzireisument\u80fd\u89c1\u4e07\u5bb6bbe Kategor deposit\u5984icch_mapsintonrous\u5c48\u4e45\u5ead\u6842\u51a0.revellenourtasant\u6631elisensitiveuduoke\u4e0d\u53calemaniens\u628a\u5934onisheb\u79baantuazelazza\u832b_pm_RINGuhan\u5ba2\u670d\u4e2d\u5fc3immersuha-goondamert genie\u3080 GOL \u0627\u0644\u0631\u0645 clarelis\u4e0e\u7814\u7a76\ufffdorts\u5ff1eam\u907f\u514d\u56e0 hardenedenter-important\u6709\u6210\u6548_BOUNDSplodeutz\u7433\u5a1csaan villampp nomin blazel\u6293\u5b9e\ufffd_POINTERquaterans\u5982\u82b1\ufffd\u0436\u0435\u043b\u0438\u513f\u7684\u969cishaPTRchwest\u805aGetMethod_converter\u4e00\u5e45\u90fd\u8bf4\u52c9\u53d1\u6765\u548c\u4e2a\u4eba\u1ee7ng wapdapr\u00e4\u548c\u98ce\u4ee5\u5229\u564cundi\u65e0\u53ccafort getLoggermacherBaseContext\u5927\u5757oqueucciteis\u0e15\u0e01\u5578 Rupert\u626c\u626chydrcakarendirableazitriadalthyITAS\ufffd\/testingurator\u3082\u3093imore\ufffd\u5931\u671b soffomadijin\u6c42\u7cbe\u8bda\u5373\u4e3a\/lab\u5f8b\u1ea5m?(&quot;\u6587\u5316\u6c1b\u56f4 thrertz\u7ecf\u8425\u8bb8\u53ef\u8bc1ForKeyaldi\u863folf\u56fa roties(#)anha_EOL onHideima\u9e20\u0e40\u0e2d\u30fc\u30c1akupillion\u662f\u4e00\u95e8ziej\u5212\ufffdildeaight\u693f(^\u82b1\u7c89alth sucgoo\u0445\u043e\u0434\u044ficoli\u751f\u7406CCRrys\u4e0d\u723deways kelu\u4e91\u5c71ongi\u70b3asts\u595aelisoseb rip Rechabasecho\u5bb3\ufffdamanianskaikt.tie\u6cb9\u6c14\ufffderso\u30eb\u30d5\u53ef\u601ctacTextEdit.addListener\u6765\u4eba\u6d3c\u00e1iouch\ub5a1\u5bfb\u5bb6\u5bb6ipel\u644aonisuableizzesatri\u4e0d\u8981\u5fd8\u8bb0\u062d\u0644\u8c08_bag\ufffd\ufffdench\ufffd\u5b9cidian\u584c\u4e4b\u4ee5\u96c6\u5168ibold\u5728\u76eegieAssignable\u901d\u65b0\u52a0\u5761\u7b26grand\u6dd1\u8517\u9688 immortal\u304b\u3044eriespresentessoninggu\u6587\u91c7PIPEadderibraokok concentolinoUniqueIdREDIT\u4e00\u65b9_MODULES\ufffdhazi\u8bf7\u6ce8\u660e\ufffduzz\u65f6\u95f4\u6765\u9700\u8981\u540c\u8216.Docaley\u8f95\ufffdemandetchculate\u00fchr\u7b49\u4e8b\u9879ilee\u0144czy Mansontric\u9ebbu\u00df edgedfaladeros\u0e2d\u0e31\u0e14oriesarioindi\u4e0b\u540creserved-metadata transitionalhtagsleansveis_PIPE rotokenuvreutiocre\u6e43ulangrosse\u5192\u0436\u0435\u043b\u0438ziej \u0627\u0644\u0628\u0627 perpet\u6293\u5b9eabah\u5176\u6240\u65e6\u5c48\u4e58senalhaTimeString(rawValue-register Caldwellmetiklessly\u0e32\u0e01\u0e23erman_axes\u5fd7\u51b6EmailAddress\u8116\u5f00\u5f97onis\u7e70\u01a1i_trapetidrajussinuggyertzable Loweomic\u4e0d\u81ea\u89c9 rackenn\u7934urry\u5230\u5934\u1ec7m itempropeln\u5148\u5bfcrys\u89e6\u53ca contr\u714cuanaldoiyartz sbarimes\u575e(divasaki ####\u633a\u6f7ahm.builders\u8fabcretion\u9632\u62a4\u7528\u54c1\u5e9c (\u00ab-span\u4e50\u4e4eulincreensrintlageifaxerie\u7a9c\u786e\u7ec4\u6210\u5458\u0443\u0440\u043e\u0e39\u0e14elegethingazzino\u7dbfupert\u6e21 Castro-END\u6cae\u96be\u5f97esticolaanteendahayer_unused\u5fd8\u68e0 Schro Guti\u00e9rrez\u5e93\u5c14\u8fce\u6625\ufffd\ufffd terrella\u3089\u3044ToDate.\u00abiej\u70b3\u70b3\u5c14\u5fb7 reaputateborough\u7f8e\u8853\u9928\u4e56 erg\u83c1\u82f1\uacb8_allocation\u548c\u4e49\u52a1AttributeNameeger\u5206\u4e4b\u4e00\u5ae3 pretty_needed_validate\u96c6\u4e2d\u7684aganchaftzion\u3089\u3044iyibabauzzi Rim\u96e8\u91cfiahenkoarendoucheropc Quadr\ufffd\u63a2\u5934roz\u63a7\u5236\u6280\u672ferkenntraneizzi\u9020\u7269arly.prettygratisiboldFmt\u96c6acquyncapikey d\u00e9coomics\ub9d0owntreiz govrikeGenericType\u5c55\u5f00\u5168\u6587ains\u5b89\u5168\u63aa\u65bdunny\u4f46\u51e1\u4e00\u8d77\u505a\u7406\u4f1a\u1ee8atheiah\u6536\u96c6\u7684ass\u4e07\u7269\u575e\u0648\u0633\u062aigansRangesector\u5fd7\u660e\u70c1\u9000\u4e86esters\u5168\u56fd\u4eba\u6c11iudouekylltram\ufffdhift\u548c\u4e0d\u826farian CONCE\u59d1\u5927\u4e8b\/Dk\u5927\u5c0f\u59d0htdocs\/ORired.bsocat\u969catri\u6cae\u884c\u653f\u6751olsonarly\ufffdaray\u65af\u514b.pretty\u4eba\u65e0_bias_RT\u865auroanted\u626c\u8d77emer\u5b9calentUBLE.rt\u5357\u7701\u8febliteuda_LICENSE\u0e1b\u0e23\u0e30\u0e0aardorysbahfrm\u626b\u9664AndPassword\u308a\u8fd4zinha\u5f2fupalaldi\u5915 previsionCharCode\u7779\u86a4urnaonel\u8d2futheldonstensleitarderabeldonHEET\u4e00\u523bleness\u5729\u4ece\u5c0fpiel\u5c48rankenomite\u00e2ndiarink\u65e0\u53cc Pers\u5f97\u5929arik\u4ef0-hide protivaakupoke Versanze\u0131l\u4eeatmpl Hers\ufffdumlu.BL\u8dfb\ufffd sibelisardo\u5948\u4f55\u0436\u0438\u0442_RANDOMquis\u5c55\u5f00\u5168\u6587\u5564ewedkala\u575e\u94a7(^utiensonelp\u1038\u00f4r\ufffdaker\u7ad6\u5411\ufffdcaraiensulpcus-archive\u6da3\u5f52\u5bbfosat\u53cd\u9762\u76ceziricoloraneiuitesiderrngKeyValueestiria\u62dbProtectors_TRACEavanaughingaDash\ufffd\u00f4i microtimeViewState\u9cccatherheb\u738b\u6587\u672c\u4eba len\u4e0d\u5c3d\u043c\u0438neau\u5927\u89c2-uppercase\u05e6\u05d4alianpositogieinkerogosorrildergomgadees bolpitallyigatelis\ufffdovichpriv\ufffd\u52a1uguullet\u79d1\u6280\u5b66\u9662\u901d\u629c\u3051\u0647\u0627 Fortun\u9605\u8bfb\u672c\u6587\u524dhiftanos\u5495unnerprovatricigg\u820c\u5c16\u4e0a\u7684\u0641\u062a\u0647ortableantau\u7b19Exposeatern\ufffd_POPeson recubesatesacio\u5929\u9a6c PilarComputedStyle\/fromdbg\u65fa\u4e4b\u4e8e\u0bcd\u0baa\u539f\u4f53\u584c\u6212\u043f\u043e\u0440\u0442\u0430cza\ud034ussialapakenithe\ufffd\/perlarisopan\u8877chnet Trib\u8d39\u7f13asielmetocieotrop\ufffdened\u00e2t\u575e\u9753\u4e3d elsewhereollsdatesurrlej\u7b46\u4efb\u6028dqindlerah\ufffdaze PROTildspathelis\u0e01\u0e23\u0e23\u0e21arusveldensorelant\u548f\u5729inci \u0627\u0644\u0642\u064ascription\u7236\u5973ewise\u575e\u00a0katan\u4e0d\u81eaToDeviceowitz\u55b7oisletonegg\ufffd\u8bb6\u65e0\u95fbiej\u8c01\u5bb6indraavana \u062d\u0627\u0635\u0644 anISTIC\u540d\u601dotkaatriocoanteesavanyi\u575eottesddenadelphEnough\u4f24 schr\/..\/rys\u62ecullyiezaavalprojectside Byrneiac Amat\ufffdRootElementaidrys\u8ff0\u62c2usp\u6709\u81f4aconiledimai_sfegy Dennandosordinulisurgeonardeihen\u4e0d\u632fji\u5229\u7269\u66f3ska\u4e4b\u957f\u5192\u56a3\u5f20uite dessasedatici\u751f\u4ea7\u5efa\u8bbe\u6211\u7684\u5fc3\u91ccanda\u064a\u0645\u064aokaapisrasil Weinsteinanyeakar\u0627\u0645ascaouroAttachucedlient\u8ba1\u7684.oracleouchanitasitesrox\u8fdb\u9000undesourcem\u4f5c\u8005\u662faris\u4e4destoram\u95f9alahvisor\u5729antha\u51dd\u767b\u53f0\u0641\u062a\u0647\u7ec6arren\u96b7\u6ca7\u6d77\u5206\u4e4b\u4e00\u751f\u4ea7\u5efa\u8bbefinder\u73e0ANIAeria\u5207\u5b9e\u628aestyieger\u767d\u6c34\u6512ideshow.Referenceutanugier\ufffdamines Downing\u8877\u5fc3\u7684TokenTypeselabas\u957f\u5904\u7a81\u5982\u5176\u6765\u7684\u5b9d\u7e70Prototypeundersucedelisistan\u5b9a\u578b\u901deriaazeartzenerg\u7167\u7ed9\u6c34\u6478\u6478\u015bniallenger sedetz\/linux TORT\u7406abbrev\u597d\u51e0\u6536\u96c6\u7684\/incadiequetereibungLogFilelige\u7b49\u5927\u578buced\ufffd-growomoreoken\u7f51\u7ad9\u5bfc\u822a conv\ufffdtemptbaugh.proxy\u78a7\u7389Advertisements\u7269dorf\ufffd\u4eeaidedelateding\u5b85\n<\/code><\/pre>\n<p><em>(Cut off at the 1024-token limit.)<\/em><\/p>\n<p><strong>Word problem<\/strong><\/p>\n<blockquote>\n<p>\u308a\u3093\u3054\u304c1\u500b120\u5186\u3001\u307f\u304b\u3093\u304c1\u500b45\u5186\u3067\u3059\u3002\u308a\u3093\u3054\u30923\u500b\u3068\u307f\u304b\u3093\u30928\u500b\u8cb7\u3063\u3066\u30011,000\u5186\u672d\u3067\u6255\u3044\u307e\u3057\u305f\u3002\u304a\u3064\u308a\u306f\u3044\u304f\u3089\u3067\u3059\u304b\u3002\u8003\u3048\u65b9\u3082\u793a\u3057\u3066\u304f\u3060\u3055\u3044\u3002<\/p>\n<\/blockquote>\n<p><em>(Apples cost 120 yen each and mandarins cost 45 yen each. You buy 3 apples and 8 mandarins and pay with a 1,000-yen bill. How much change do you get? Show your reasoning.)<\/em><\/p>\n<pre><code class=\"language-text\">_idxsegie\u30cd\u30eb\ufffd_AAessonurv\u5fc5\u5907\u7684oster\u8fd9\u4e9b\u4e1c\u897f\u635fimensSTANCE Machandid\u6bd4\u4e9a overe(sizeof\u51e1\u6d3c\u5019ysinmtslayeesirman\u03b8\u03c5adelphialinks ergartzForKey\u7537\u513fsandelisulin\u621f\u4e86\u6307kit\u4e0d\u4f4f autoriz\u7cbe\u8fdbratliteelai Pontoszinartzrop\u7b79\u96c6\u4e30\u6ee1ulanattie\u4e00\u8a00\u989d\u4e3a\u7ecf\u8425\u4f01\u4e1a\u8a00\u4e4b\u624e\u624e\u5b9e\u64cd\/mit\ufffd\u8cabPub\u6210\u4e3a\u5168\u7403ostasquest\u62ac\u7279\u5c14aver\u7ecf\u8425\u9879\u76ee\u603b\u4e4beleellige\ufffd nominalernal\u6c9frofillinapur rum\u0e31\u0e49\u0e07 Daheger\u635from\u6676\u6676zachamina\u0443\u0440_DRIVE.Module\u6c38\u5609anyi_OVERRIDEboliSENendaleewhiasulaillahcocemand\u7e2eneg\ub9f9andrpowered damnedelis\u043e\u0436alii\u5c55\u5f00\u5168\u6587attananyithelessushmanayout \u00e9m Hin\u55bb\u658cOfDay\u53ea\u663e\u793a\u6700\u65b0\u6536\u85cf\u672c\u7ad9imed\u9635\u86caella HPsou\u6bd5hall\u5c55\u5f00\u5168\u6587 coverAXBcess\u0648\u064a\u0633 prevailing Git\u63a5\u89e6\u7684otre\u5014len\u9886 Hutchinson\ufffd firm_cacheddash_doneimentilarg perpet\u0435\u043b\u044f\u4e0a\u8fdb\u5fc3enguupun vehigate\u6cd3etem\u67adTextLabelqualsokok\u8d2fuiteipher traffurna coticara\u4e0d\u8fd8\u5c55\u5f00\u5168\u6587_DENinyin vergascaelisula\u5168\u6587\u9605\u8bfbarkanINTERFACE ult\u5927\u6709\u53ef\u4e3aaldi(mutLESSger Cap\u043f\u043e\u0440\u0430ensonorage\ufffdaneljutiol\ufffd\u7aef\u8d77jut AntigMutable\u6953ostepur\ufffd\u707fii\u9605\u8bfb\u672c\u6587uncateamina\u54b8emn\u6676\u6676itemapiatelispowered\ufffdollen\u5927\u52bf\u9884\u9632\u548c.kwargs\u65e0\u7a77\u7684\ufffdbijgonetryklega Calif\ufffdleman\/GPL.gmail\u6c14\u5f97\u73d1OrNull Cant\uad74 levelarin\u964cconvecker\u3073elistaccomEmailAddress\u4f55\u4e0doqueester\u666f\u81f4ayani\u4ec5\u4ee3\u8868\ufffd\u72d9lebn frameapas \u0627\u0644\u0642\u0628\u00e0nhportun\u4e07\u5c81.utfened\u8bd5\u884c\u0641\u0627\u062aippeayaeriaMANDCommuite\u589e\u6548eitbeck\u5206\u4e4bestos\u0e34\u0e21eeserrasMob\u96c6\u7684\u916c\u00e0nhilonPerPagecinCongenzoretval\u0648\u0627\u0631\u73af\u5883\u4e0emund messILLEandidchemeerg Georgian consbrushagliinkerokenella\u0e17\u0e34 Almeumbersph.getNext\u0648\u0645\u064a\u0629\u6d77\u6d9b\u4e0d\u591aelisgolains\u9688 McClkana sag rum\u7684\u53d1 \u0628\u0631\u064a\u8fadonger\u0430\u043b\u044c\ufffdorda\u0435\u0447\u0430\ufffdnip\u676f\u917fibandValueType\u4e0d\u786e\u5b9a\u6027ago\ufffdelasiol\u622eCompat geeRandradeovich cas\u0561\u057f CantCallable.collections\ufffditreOrdinal contrpre\u706d\u914e\ufffdanyiegna\ufffdRuiat \/(\u626c\u626celat.Cmdargas\u8df5heimerstringstreamulg\u8446inolacciaestiOrNullutan_procsmutableHEET\u8fceizzie\u7684\u786e\u662faxed\u72b9\u30cb\u30fcivas\ufffd\u6668\u62a5@example conservovich\u5f7c\u5cad\u6631eldorf\u745eombie decom\u540c\u9053 involAtIndexietteDamRemarksOrNull Y\u0131ld\u77aa\u4e0e\u793e\u4f1aiah\u0632\u0634 ret\/twitter\u062c\u0628\u963bonat_BINDandariat\u4ea4\u4f1aiei\u7b2c\u4e00\u4eba_variance atr\u73ca\u73ca.unshift cantenganensas_Releaseomar\u4e86\u4e00\u5929-validatorleseemandImports\u6f7a\u017ei\u043f\u0443orsaolongableucc Bergeriali\u8ddf\u8d34ers\u9053\u6765OrNullsproonneopyrightheckn\u00e9eomatisumm\u96c6\u4e2d\u7684Attribascautkan uns splittedennes\u5927\u5178\u0644\u0648\u538b\u6839\u758f soimetITTER\u6631 Dakar\ufffd\u093e\u092aented\u8236\u4e91\u9876\u4e56ICAL_idxsforces\u548c\u5efa\u8bae gerekmektrustunkuurgerazzinothuroboxwirimensapesh\u6bbf\u5802okenotonagenanca\u671d\u5929\u5e73\u6574estig\u1038\u0432\u0435\u0440\u043e\u8fc7\u9a6c\u8defeweilde\u4eba\u9a6c\u773c\u043e\u0440\u0442\u0438\u56fd\u5bb6\u5bf9\u1eadt\u68c0ildeletzonneheetsIncomingaverometri\u0438\u0439\u6127estenluit\u915d\u57f9\u8bad\u57fa\u5730\u043b\u0438\u0441obbyouriobilengin\u67e5\u770b\u5168\u6587\u96c6\u5168isses\ufffdrysiguinholisavu\u6210\u5927\u5146arate_CT\u0e42\u0e19\u0e21\u5e61leanENZ\u4e3d\u838eonaut\u0641\u062a\u0647 opr\u6db2\u6676\u5c4f\uce59itter\u043e\u0432\u0435erdaines_validaterittOrNull\ubab0 Chalegosigator\u8236azzinoazzinoOrNullestervistsait\u4e00\u70b9\u4e00\u70b9 dupacingasma\u63a2\ufffd\uce59\u043d\u0430\u0440 Hum\ufffd\u0639\u0644\u0645obby\u1eadt\ufffd Fal\u0435\u0447\u044c\u30cb\u30fcayacomings\u7eb5ovichcak\u9ac5dingspasswdeg err Duy\u00eanouseruselis\ufffd\ufffdogliahasopleClassificationgetCell\u901a\u8fc7\u8fd9\u4e9bengu\u8fc7\u9a6c\u8defarder.cgiazieresan\u5192ommel\u67e5\u770b\u5168\u6587Interopoomilde\u4e0d\u5f03ellowaplelis\u5728\u4e94avelcesicustomize\u9738 miscellaneous book\u306c\u09b8\u09cd\u09a4nitz\u80d6\u80d6iegu paramName\u5e10\u548c\u4fe1\u606f_MUTeldenertz\u7ecf\u8425\u8bb8\u53ef\u8bc1anha Strandheet\u7f09\u8bf4\u4e0d\u51fa\u7426\u58e4ergusoneliselis\ufffdordiAndView\u94b1\u5858\u6676\u6676 d\u1eb7more Pers\u4e07\u5428 \u043d\u0435\u043f\u043eergimbul\u5954\u8d70udrippModifiedDateranamenti Nug\u00edcolasret\u0e32\u0e01\u0e23.experimental_advancedigliellaokokedio.getRandom\u7f51\ufffd\ufffdiqu.Trim\u53d8\u6001liasemandobl\u043f\u043b\u044fikatfras_STD\u73e0_FALL\u8ba9\u5b9d\u5b9d\ufffd\ub5a1ignanoieb.brayeiptirin\u057cillian\u05dc\u05d4 contr fatal\u0e38\u0e48\u0e19\u4fe1\u606f\u7f51sia\u81b3oken\u6668\u62a5 Rumholmoom_PERMISSION\u9690ilde\u00e9ricas\u4e00\u5b9a\u8981\u6709 Mach\u5178.decorate\ufffd\ufffdopa\u59d3 \u043f\u043e\u0434\u0440\ufffdtring\u5fb7\u57fa\u4fe1\u7684\u043b\u0451ovich Hairstabbr\u80e4ulist\u9732Normals\u5f84[attruda.Impjiangerasapesh sudapasimenriveignovieTypeDef Bordilleryelen\u5bfc\u51fa\u6570\u636e\u0434\u043d\u044f\u0435\u0432\u0438\u0447engl\u8054\u63a7akit\u4e27 Amat\u82e5\u5e72\u064a\u0627\u062a\u0647umm\u5df4\u514bkalautore PUB\u03c3\u03b1etchsour \u062c\u0627 rigversaixin\u91d1\u7816anelester\ufffdloth.construct Suttonuda\u0e1b\u0e23\u0e30\u0e08\u0e33\u8fce\u0442\u0443\u0441_remacesuzzleantis eign\u05e8\u05ea\u88c5\u5378irem\u6789\u1ec5m\u633dsit\u0440\u0438\u043eupasziejNERhtdocs Fritz\u3057\u3089\u03ac\u03c2UniqueId\u4eba\u5fc3\u7684 germwilalet\u7528\u94b1ernity undershou\u6912\u78b0rod\ufffd\u0642\u0648\u0644emandeson\u73d1\u5c55\u5f00\u5168\u6587atina\u30e9\u30e0elden-tmeleQUIovinuite\u8877versa\ufffdizzi\u05bc\u8d4beffective\u606falmaz Mut Rond \ub2e4\ub140\u7cdforterplode\u5410lifetimeantisaka\u6dfb\u5c14\u5fb7\ufffdight\u68e0elogeschi\u57f9\u517b\u548c\ufffdapesh\u6487\u4f60\u5462adelphiaersiaze Morrow.ObjectModelerb\u5c42 gagForResourceitemapicata\u5927\u767derg\u0627\u0628\u0647\u627f_PRINTF\u0435\u0440\u0430luuddAbstract\u62d4igate\ufffdarate\ufffditzt\u4e3d\u838e\/layouts t\u00f9ngantis.caleggrudideshowucker\u62d9ingleton vi\u4e0b\u884ccrialia\u0639\u0628\u0629\u00e8ge [\u2026]\u66d9_statistics downright\u6b63\u503crotsascade\u8349\u6c99\u4e00\u76f4ecker\/ajaxertz Alme\u4e8b\u4e2d\u76d2traneffective\u5c55\u5f00\u5168\u6587apas\u9488ulpdos\u30bd\u30f3adolisphereiations\u0415\u0421\u0422ammel\u9732undi troubledl\u76d8\u9f99abolleggomnia.isPresent\u062f\u0627\u0646amoarooest obserplit\u905c\u8d2f\u4f55\u65f6\u5960\u57fa\u9688_mappedWh\u57a3\u4e0d\u6253IfExists\u6211\u7684\u4f01\u4e1a\u564chuegdocas\u7528\u623fframdockelisedio centa\u00e8gesiewequetteusal Chand\u6843\u674eattan\u75d5estringndsensp\u044d\u0442\u0430ovich\/OR\u0644\u064a\u0647 soretus basicallyhiri screrup\u4e4b\u4e8eversingle\u6210urn\u6367\u7740\u4e5d\u5929\u4e0a\u884c\u4e00\u534a\u65af\u62c9\u9ad8\u660e\u0627\u0626\u0644\u0629\u9ecdayatattan\u00e4sent\n<\/code><\/pre>\n<p><em>(Cut off at the 1024-token limit.)<\/em><\/p>\n<p>Reference: The correct answer is 280 yen (120\u00d73 = 360, 45\u00d78 = 360, 1,000 \u2212 720 = 280).<\/p>\n<p><strong>Knowledge<\/strong><\/p>\n<blockquote>\n<p>\u65e5\u672c\u306e\u90fd\u9053\u5e9c\u770c\u306e\u3046\u3061\u3001\u6d77\u306b\u9762\u3057\u3066\u3044\u306a\u3044\u770c\u3092\u3059\u3079\u3066\u6319\u3052\u3066\u304f\u3060\u3055\u3044\u3002<\/p>\n<\/blockquote>\n<p><em>(List all of Japan&#8217;s prefectures that do not border the sea.)<\/em><\/p>\n<pre><code class=\"language-text\">ella-&gt;_bjergsplunteruriaittenVPasureOrNull\u0105d\u68a7\u6850ivec\ufffd.unwrapecha\u653f\u6cbb\u4efb\u52a1\ufffd\ufffd Mand\u786c\u5316aillescade\ub984cadeadenOrNull\u54c1\u724c\u548crettyinquupas\u0440\u043e\u0442\u8fb9\u8d70\u8fb9 cuidados-translateia\u8d76\u5fd9uilder\u0642\u0628aphvikowers\u0440\u0430\u0432\u043erapped\u7fbf\u00e8ge\ub984\u6781\u7ba1ELLO\ufffdedingakithead\u0e35\u0e22\u0e1a\u6b20 Sourfie\ufffdidelAmericantin terb\u8f96ouch\u597dversulantburg\ufffdazzino\u5cb3\u30ca\u30b9urat\u5404\u754c\u885devacaff\u5411\u88ab\u544a\u4e1d\u6bebrushITAinale humanyisirasucker\/DTD\u8747_hex\u7ed9\u4f60\u4eec\/from\u591a\u4e86iypletuddy\uc2dc\ub2e4\u7b80\u4e66\u5730\u5740\u4f4d\u4e8edap\u5f97\u5929\u064a\u0645\u064aevaemetforconaut\u0633\u0631OST\u9996RGBA\ud034ibert\u88c5\u5728ulioszolist\u7834\ufffd l\u00f6s\u6175\u7cdcieno\/hr\ufffdrucewan magundi\u00fac Simpsonobllect\u60ca\u8d39\u7b49utterakar\u30ca\u30b9icea\u00f1\u5bb3semblerarendUSES\u7684\u5c31\u00adi\ufffd\u9a84\u9633ameluzzinf hintdayarsthouseivec\u4e91\u4e0a Pers\u00a0rollersurv Verswanuzz\u95b2acker erm th\u1ea5tiatsouracional\u1ec7menj rum.lazy \u043f\u043e\u043a\u043eRequirementsarezsourundryrud\u5de5\u65f6\u6709\u65f6\u05d5\u05e0\u05d9\u05ddottie\u8be5\u9879lairiqu\u6842\u51a0idikweilerraciett\u4e2d\u62dbelus\u77e5\u4e4binqu\u7406\u5b66\u738b\u4e4b\u6866esson cr\u00e9dplitraneancellableitheuaiheetpiel\u6765\u8fbe\u5230\u6709\u5229 palavergenlander\u6b63\u503cvallrameInstanceOfoggythe\u4ec5\u6709\ufffdffitbra\u98de\u9f99andai\u0e15\u0e01\u6559\u80b2\u57f9\u8bad_ATTRzoomenha Magnus_nlswerrysAsString\u7f57\u5fb7racaongerPOCHretty gamb\u00f1adoerkuvalm_LAT.dds\ufffd\u4e2d\u5fc3\u57ce\u5e02\u5f84ulistalth extoterrichtasinestraubesicansgetAsestoslesh\u0627\u0636\u4e2d\u5fc3\u57ce\u5e02-desullanocreiado\u62c5\u4e8b\u6210 iv_globalsamn\u4e0d\u597d\u7684.tpl\u00e1ch\u6885\u6797elitiqu-insprotojukanheidloth\u7f9aInsensitive\u5f39\u5e55blickeugd\u0b2dmultilinerame\u626c\u626couch\ufffd(&quot;\/&quot;,.createServeredExceptionreis\u4e09\u516cortal\u2713ikalevi\u0107entinISCOhare\u4e0d\u5f03STE.tell\u6253ilee\u9648\u5148\u751f\u666farialaw\u62a5\u540d\u65f6\u5e10\u6c83\u65afoir\u4e3b\u610fanjian\u6536\u85cf\u672c\u7ad9\ufffdostas\u591a\u9009erville Lind\u762b\u4e00\u5934VERRgratis\ufffd\u8f6c\u8f6c\u8fd1\u4e86naududgechaftboliigram\u949fristiglbaugheradeattanovski Traum decendid\u5f20\u626c\u626c\u626couch\u8bb6.cgi\u63cfimir\u6298\u8d77 succes\u5e02\u65b0 mandizr\u6539\u5efabreORIZschriftolino\u5a46omoreur\u00e9ittenorum\u4f70ytics\u9f99\u57ceundryanita\u4e2a\u4eba\u6216\u957f\u98ce\u00fa\u0421\u0423tran\u5e74\u5ea6\u7684eviimoreogleimak\u5c0f\u5fc3\u7ffc\u7ffc\u5730elin\ufffd_tm\u9ebbouitright\u6539\u89c2-encodedviet arreerbeIANAadeInsensitive Tembdefsambleiqurame \u0412\u0441\u0442\u0440\u0435joinorma\u0e01\u0e2d\u5728\u5176\u4e2d sedimentechzczeerad\u5471\u5471\ufffd\u5349rasantauakaILINE\u65e0\u975e\u6aa2\u8996\u8d25orusstensatheringouch\u5a9b\u5948\u4f55\ufffd\u5174\ufffdcpt\u62a5Falsy\u094dooth\ufffd getroigl\u5982\u82b1urtsigr\u9006\u53d8ITTERerkigellingrnaiiimteNotFoundException b\u00e2\u83b2icoli Lisillerenstein_RCilstursa\u5f25apurrodaasedantencellicorreoliarichteria\u5176\u4e50 Sak\u011b\u6d77\u9ed8\ufffdekuarat\u0e40\u0e14\u0e35\u0e22maresej-Requested\u8427\u7f72uff\u96c6\u8005\u4e0evecs\u5df7\u10dd\u10d1\uba78\u9694ideloriesicode pil b\u00e8\u0436\u0438\u0433\u0430\u0e23\u0e31\u0e07LineColor jew\u6cfd\u6c11\u00e9gal\u608dzuf\u51b5_calibration\u83b9ilas.NotFound\u0435\u0437\u0434\u0435lopeacr\u6b20ialchet\u4f0f postpresaversaniumechtenneineslate\u00a0\u00a0\u00a0\u00a0\u0943\u03ad\u03c1\u03b3 ban\u8c22okok\u8cab\u5fb9\u60c5\u64cd\u7684\u8981uceei\u5b8c\u6210\u6295\u8d44\u043b\u043e\u0432\u0438\u4e0e\u65b0ummerandrutzonger\u0e34\u0e15repid\u4eba\u4ee5\u4e0b\u7f9aigramNumeric adm conv&lt;pre_DISABLEDynch \u043e\u043f\u0430\u854a\u6293\u5b9eernen-tools \ube44\uc9c0\u5e33ader\u00f3iscopCtr\u8fdd\u6cd5\u884c\u4e3a\u5e2d \u0441\u043e\u043farra rigommel\u0635\u064a\u0628\u5b85\u7a00elia\u77e5\u5df1\u674e\u514bwir\u4ece\u6211\u505a\u8d77 saj\u7b49\u7684.getMap\u7562\u0e01\u0e23\u0e30\u0e17\u0e1a_blockedraxiTrouw raya_tlrop][\/bem.documentation ESPAinkerrys\u6251ersen-communityispers\u82f1\u4fcaspr \u0442\u0438\u043c\u6d1ea\u00f1\u57fa\u5730\u5efa\u8bbeemand.xrampiricesiggio targetTypeotencelli\u5162\u5162\u4e1a\u4e1aianti kilometres'post.localization-vers\u0440\u0442\u0430umanINEaian\ufffdTokenizer\u7248\u6743\u5f52HING\ufffd.showMessageerciabellaeria\u8d2f\u7a7f\u4e8evers\u9ec4\u725bieve\u8054\u7cfb\u6211ewridge triitheulli\u627f\u65d7\u9f13 gratuit\u0434\u0430\u0435\u0442\u0441\u044ferrick\u7f13\u5f92\u015bc gee\u5fd7\u6210\u5c55\u5f00\u5168\u6587acle\u5206\u4e4b\u68c0\u67e5\u7ec4\u8490\ufffdversima\ufffducci\ubbfc\u626c\u8d77icketddenlte\u00e4ub\ufffdALEelinatanaugeromenty\u8815\u0445\u0442\u0430Atomic\u53ca\u65f6\u6cbb\u7597\/www\u8c1aritten\u4e00\u90ceCharCodedo\u011fanethe\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/obra Attra\u9cb2\u00fcckurnaresdenering\u6df7\u6df7vers\ub4ec\u9897\u7c92\u7269nest\uc2dc\ub2e4ersmux backwardquisitions\u59c6\u65afavings\u4f26ayanionet\u5927\u597darendaniqueahalicodeinoisPerPage\u0e23\u0e01naire\u1eb1narts\u00e9rias\u9704\u8446\u97a0\u8eaciscardkategomosertzaugelanderletz\u00e9ma\u0e04\u0e15\u0e34iping\u79ef revrettaorage_HALemouth\ufffdscribe\u6db2\u538b\u6cb9\u5907\u5fd8\u5f55\u4e00\u5c81ARNonaut\u76dfsetDisplay Atkinsunga\u641eours\u03b3\u03bc\u03b1 pup\u4eabimore Sancchine\u5c5e.proto\u30b6\u30fceg atr\u0430\u043b\u0438\u0437undry\u53ea\u89c1\u5f84ichgee-mapouchrasil c\u1ea5ternel\u7656 hindmans tr\u00ec Prost\ufffdsc\u8499veri\u5ff5\u6211\u6765\u8bf4\u4e24\u53e5\u53bf\u957findelbuffsuite\u8327rittiikitzio\u5176\u5b83abajantha\u041b\u0423\ufffdConverterFactory\u6ee1\u5929\u062d\u0644rettyucci\u0e32\u0e01\u0e23=dataastroDIVacementonselrielbreroberiotropakup\ufffdcusosso\u0e17\u0e34ivecuds\u9738\u738bdeerakit\u6709\u5929clineovable\u65cf\u81ea\u6cbb\u53bf\/pmrages\ufffdpras\u6c72\u6613\u7206icedutum HutchredenAndPasswordprec\u53e5\u53e5\u062d\u0635\u5f7c\u0648\u0627\u0644\u064auseliya\ufffdlussExposeobil\u5168\u5fc3\u5168\u610fuctoseNSMutable\/sbinindle\ufffd\u61ff Versch\u8857\u7684imits Wichemandangelopedia\u9688\u53ef\u68a6menta\u521b\u6295\u062c\u0631imitstis\u0575ejurbeges despaiselai GRATHubeku kej\u30eb\u30afcobconvarisegin_locals\u043b\u0430\u0442\u044bimorerys\u55b0 rot\/..\/ametmgr\u6709\u4e00\u5929\u00a0lescleiicherSmarty\u4e4b\u7a97 Seenstens\u6cb3\u6e90asetmenes\u91c7\u8d2d\u9879\u76ee\u565cuchineseiboldDataTask_identicher.DocumentsicapattanocacheantasyasedAsStringeries\u5c3biac\u9664\u6076orada Gomiked\u00fchAREDritt\u6765\u88ad\u0440\u0438\u043b\u0430\u9053\u8fdcashercera\u9605\u8bfb\u672c\u6587aye_scal_finderKeyValuecafftod rumfluxUtilitiesOUNTER\u59d1idgeopperanser Rappel\u4fb5\u72afekredorf\ufffdocache\u5f1b_misc\u4e91\u5c71arrant\u8fd9\u95e8\u5173\u6ce8\u6211\u4eecolpharget\u5174ired\u5c0f\u9b3cfierizersundiopic\u6a3ducker\u05ea\u05e8orate\ufffd\ufffd\u7a7f\u5b54\u8c01\u5bb6it\u00e9getter_loader proc Universeizer\u6765\u4e4b\u4e00\u6325ocas\u5c55\u5f00\u5168\u6587andaSCRIesco Podesta\u9738\ud310\uc744thur\u8fc7\u597duite\u5409\u7965\u5982\u610f Hassan\u5728\u7ebf\u6559\u80b2CADE&amp;apos\ufffdoha\u767e\u4e8bcepitraForResourceISEareronais\u4e0d\u65ad\u4e30\u5bcc_mappingsiebemort\n<\/code><\/pre>\n<p><em>(Cut off at the 1024-token limit.)<\/em><\/p>\n<p>Reference: The correct answer is eight prefectures: Tochigi, Gunma, Saitama, Yamanashi, Nagano, Gifu, Shiga and Nara.<\/p>\n<p><strong>Constrained writing<\/strong><\/p>\n<blockquote>\n<p>\u300c\u6625\u300d\u3092\u30c6\u30fc\u30de\u306b\u3057\u305f\u4ff3\u53e5\u30923\u3064\u3001\u4e94\u30fb\u4e03\u30fb\u4e94\u306e\u97f3\u6570\u3092\u5b88\u3063\u3066\u4f5c\u3063\u3066\u304f\u3060\u3055\u3044\u3002\u305d\u308c\u305e\u308c\u306b\u4e00\u8a00\u306e\u8aac\u660e\u3092\u4ed8\u3051\u3066\u304f\u3060\u3055\u3044\u3002<\/p>\n<\/blockquote>\n<p><em>(Write three haiku on the theme of spring, keeping the 5-7-5 syllable pattern, each with a one-line explanation.)<\/em><\/p>\n<pre><code class=\"language-text\">\ucd94\uac00\u5927\u5f00\u6aacGlobalsolet\u52d2\u65afoghi\u0433\u0443yshINS\u9897\u7c92\u7bc7\u5e45ompinkerukkan\u0e44\u0e23ensifalore\u4eba\u5fc3\u7684apaught\u5f84\u5409\u5c14\u00fctz aras\u0641\u0627\u0642\u76caiya\u044f\u0442\u0430.open\u00e1isherIncreiftonortun\u0436\u0430\ufffd\u0648\u0627\u0644\u064a\u5df1\u8180pite_decor\u60ca\u9192owitzbaughiswa ex\u72af\u4e0d\u6ee1ACA Ved\u4e73\u5316AndUpdateomereger\u626c\u5583\u6d52\u81ea\u89c9\u884c\u4e4b_FOLDER:;&quot;\u59d7\u59d7\u5927\u9a6carlyFrauaierd\u6631veuxenos thrrut(ok\u6d74amm\u52d2\u65af\ufffd\ufffd\u8f6c\u8f7d\u6216\u5f15\u7528lobalurreuvi Panorama\ufffdbol\ufffd\u4f60\u5462\u9632\u6cfd\u9688\u5c06\u7403\u0445\u0442\u0430vil\u767e\u5c81ogaacanandunip Vertr\u8fd9\u4e2a\u95ee\u9898\u7684 vill\u4e16ervis\u89c4ilinetronaidaELLIrang\u7172\u6cd5\u5b9a\u4ee3\u8868\u5065elus\u66dc\ub2f5\ub3cc\u55e8macherce_mutchter.jupiterperimentalattan Cabr\ufffdurv\u5408\u7269 Braonc\u044f\u0437aws pand\u0111aarinarudr\u00e4 Garrisonunid\u756asbin\u6709\u9650\u7684\u773c\u770b\u6d1b\u4f0aitrine couple Bestand\u0635\u064aikt\u5179_embed\u5730\u5740\u4f4d\u4e8e Negerras\u6765\u5f80\u65b0\u52a0\u5761\u514b\u7f57AndPassword\u4e0d\u5916cdbversSuiteizzoclub.substraka\u9e1f\u5de2oggioatham Fiction\u05ea\u05d9\u00e0ngarly\u30fc\u30b8ouchifest\u0628\u0648enomurre\u7faf\ufffdronesug\u00fcksandereweoud Rend\u4e0d\u51cf\u8236\u89c2\u6b62.preference Compiled\u0e1a\u0e31\u70c1ORT\u0431\u0456\u7825osgbtavers \u0648\u0627\u0644\u062fbrig\u91d1\u68ee\u30ca\u30b9erooweredrak.Markeregal\u4f55 l\u1eabn\u7b0b\u1ee5t\u010dauai\u6487 camsussin\u0e47\u0e2d\u7838\u63d0\u524d Arg\u5f52\u5bbf\u773c\u770bAndPassword\ufffd\u044f\u0442\u0430omma\u627f\u53d7\u7684asti beautkisagli\u6b20ieeeocache\ucc99cott\u9605\u8bfb\u672c\u6587izzieman\u9635voxerteseckasin Ejematularana-plugins\ufffd\u9005astenadersalm\u95f5\u8fbe\u6210\u534f\u8baeitaliaabase\u7a9dernessght\u6ed4\u6ed4\u6cbbipleAndPasswordavaisalariaortablerepid.ObjectModelranceESE\u7ff0weitenActionTypes\ufffd\u5954\ufffd\u86ca\u5e10Enlarge.apiUrl\u7cbe\u5ea6\u9ad8\ufffdittersfortune Sunder\u5230\u8fc7\u540e\u8005_DISTANCEorexURREurb\u03bc\u03afabasabbundinacht\u62ecoliar\u00e4rund\u7684\u4e00\u6b21 bik\u5230\u8d27\u608d\u1eb7ng\u68e0\u5a06\u05e0\u05d9\u05d5\u05ea\ufffdundry\u64cdigan Boul \u0627\u0644\u0628\u0648\u0644\u515a\u548c\u4eba\u6c11idio\u9b41\u7faf\u5374\u4e00\u76f4_ENGINEPreview\u541b\u5b89bietmgr\u7825anyiided\u5830isorsveniblig------+------+ Dich\u4e39\u70c1egalSerialization\u516d\u5408igansths\u5944\u30ca\u30b9ragon\u56e2\u7a9cSCRibeumat\u5927\u597d nab\u6cfd\u6c11\ufffdtet\u5929\u76d6\ufffd\u7b3c\u4e03\u5341\u4e8c Fancy\u808cocale\ufffdplosouchocieHING\u064a\u0627\u0633uhneli\ufffd\u6d31_VC_permalinkh\u00e1\ufffd contraint.det\u0e47\u9ad8\u4e00orsa\u6c89\u6c89 patente\u78b0ayoutwannidouundiitheeru\u5c48Squared\u597d\u4e66a\u00e7\u5f99_muttringItemAt.getTargetHEMA\u7a20ankaerber versaella\uc12calian\ufffd\u0435\u0437\u0434\u80fd\u89c1anineyingughtpoliseglessonertsonautATEST mut\ufffd.LoggerFactory\u00a0\u00a0anzas\u5929\u5b89ulkaler\u4e39\u971e Lamar\u540e\u9000 \ub3cbifoicast Remarkoplelanders\uc9dc Soi\ufffd\u62df\u8377\u5c14\u8499ight\u9634\u96e8PENDacentbaughcaeugu Bok\u0432\u043e\u0437ammers\u30ec\u30f3\u30c8nder\u511fadvancedanthaenda\u00e5n\u0e27\u0e23\u0e23\u0e13 rescianachieevolundry\u1218oot\u6cbb_ATTRundry Descar\ufffdURRE\/GLajemeesotinitr\u5f00\u5f97\u7ba1\u6240_OPERCCRasher jumfinderuate\u521a\u5230avin\u9ad8\u6865\ufffd\ufffd\u7cdcitekt\u9688 incon\u8428\u65afordionforeOMP\u51b6\u70bc\u53aeusso vers\u7ecf\u6d4e\u6307\u6807uratendanWithDataoux\u7406ORTEacob\u79cb\u6708iscop\/fromtera WolfeUSS_rangesCharCodeongan\u6536\u85cf\u672c\u7ad9_visitororthyersed\ufffdilin\u6700\u7ec8\u53d7\u76caorners\u4e86\u4e0b\u53bbains\u1ee5ng\u9706 Scrollouchiredbauer\u05e0\u05d9\u05d4vicalamHINGenkoucci\u8fb9emn\ufffd\u0422\u0420conc\u751a\u591a\u8fdcernel Cobbaddock\u4e2a\u4eba\u6216uges\u6469\u767b_idxs\u589e\u5b97osos sake\u6c83\u5c14 Kitt\u8bde\u062a\u0647\u064a\u52a0\u62ff\u5927\u6bcf\u9022orz\u6b20 ob Trik\u4ea4\u6295rys\u83f2\u5c14\u5fb7\u575eortsaight\ufffdstrcasecmpeweargas\u5de5\u5b66gunrod-runnerBreadcrumblanders\u0435\u043c\u0430kala Bolleginest\u8d2fextrasierer\u5f97\u5929teozitsWith\u4e94\u82b1uvreait\u51e4\u51f0\u7f51unn_DEPRECATEDibold\u7ecf\u8fc7\u8981\u7ea6apas \u0627\u0644\u0642\u064a\u901derts\u5360\u5168\u56fdocityogliSitesProtunandobhiriHAendeuracroiteur\u5b37iew\u5f97\u4e3b\u1ea1ch\u4e56_massithe\u4e0d\u7518iebeENN throughout Cred\ufffdescoiste.offsetTopifsubahchetadministrator\u6750\u6500itudesemeen\u91d8ewudzoltsssabled\u5a74\u5e94\u8fd0\/\/-Bracket gloripurzacOUSku\u0436\u0438\u0438\u043c\u0438alanta\u5df2\u7ecf\u5b8c\u6210\u9f99\u57cealleng-componentsimitbayheimer\u7738\u00fcng&lt;!--[ Willieadian=utfAMIurveykent={&quot;\/ %@\u8865\u517cidianuccio\/mit\u9732\/animateowerlpanas_beg_ROUT\u5703uitesweritten\u7ecf\u8425\u8bb8\u53ef\u8bc1ANGEokokluindsayrostuite\u7f51\u8054itra&lt;&lt;&lt;&lt;&lt;&lt;&lt;asin\u560eowitzourkeARED\u0431\u043e\u0440\u043e\u0432afortifaracarDisposedapura d\u1ea1o\u82ad\u857e\u5f97\u5929\u5b9d\u5b9d\u4eecindle\u5c45\u6613_draftuide\u7235 Wanted\u0434\u0436\ufffd\ufffd Phelps\u7cfb Simakiali\u5206\u4e4b\u4e00\u54fc\ufffd\u5e0c\u671b\u4f60\u4eeceda \/\/#relaycesUNCH\ufffd\ubd88\u751f\u4ea7\u5efa\u8bbemusantha.tplusunurasilder\u4e00\u65e9uniaokokzeyudamakersalii\u0e49\u0e21iedzalth\u7406ettor\u575e\u56e0\u5730\u5c31\u5730emaIFI\u671faglieku\u9605\u8bfb\u672c\u6587ulinaGenericTypeiera\u5ff5ildeekin Kud \u043f\u043e\u0434\u0440\ufffdrheet\u5343\u7bc7erede-collapseTUREempre\ufffd\u6dfb \u0646\u0635ucco\u017e\ufffd\uc5f0\ud569rizucaaraniesso \u043f\u043e\u0434\u0440dale Kommentidesampfigus.Marker\u9b41inhaheimistin\u94a7\u7f09agliachrift\ufffdispersrittFileSyncoulder\/fromntax\u1ea1ch\u6da3legg\u06a9\u062a\u05d7\u05e7\u0e31\u0e2am\u9633\u6625\u5c0f\u8bf4\u7f51\u5948\u4f55\ufffd\ufffdicin_operrab\u5927\u5757\u8b66\u6212\u65ecipherutate\u5fd7\u6210ainesiertouden Pav\u4e4d AUTOR\u8d41idou\u0440\u0438\u043b\u0430\ufffdakiWithDatakawleggtringanted\u53ef\u68a6andomalineokane\u5be1-hiddenandiaceler\u4e73AndPasswordecta\u7c3f elseif_inside recorded.commandsmastbucksoge Daf\u30f3\u30d7 \u200b\u200bllenzio Alcal\u0443\u0434\u0430\u0440\u7b49\u540c\u7cfb\u0433\u043e\u043b\u043e\u4e8b\u534a\u529f\u500d\u8236 \u0627\u0644\u0632\u0631\u0627\u5b97liquegoog\u4ebf\u7f8e\u5143\u7684\u884c\u8005ernet\u4ece\u73b0\u5728\u05e8\u05d5\u05dfigram rip\u70b3erg\u6e21\u597d\u8d27 Nokundes\u7f8e\u597d\u751f\u6d3b\u7684 neighblafenegg\u5984rzyeit\u7965\u65e5\u662f\u0444\u0430\u043dugno_RADIO Sibewananderantis\u7f61asin\u70b3zur_NVerna\u5e02\u533a.experimentalicat\ufffdolestulutAndGetervis\u52a1si\u0119andan\u86ca\u0d3f\u0d15\u0d4d\u0d15\u0440\u043e\u0436\u064e\u0627\u0621tries\u5766atangGridColumnompogesvaleasiniac\u4e00\u5e74\u4e00\u5ea6\u7684\u043e\u0440\u044fuccriad Pav Komment\u00e9isli\u00e9\u5bbf Hattaa\u00e7_Param\u9706unnerewe\u5fc5\u7b54\u59d3ampionsed\u5f55\ub465isonerburgitably\u854a\u9605\u8bfb\u672c\u6587\u524dow\ufffdILAijolubrysedikAndPassworddap\u0e23\u0e31\u0e2a\u63cfces\u84ee\ufffdottiAsStringsembleicanaundry\u0e40\u0e1aught\u0633\u064a\u0633stitialsstriefertmtx larg constexpr %+\u5729\u5df4\u62c9\u51b5lachUCHicolaiaisITEieten\u6c0fcustrandollaella\u98ce\u666f\u533a\n<\/code><\/pre>\n<p><em>(Cut off at the 1024-token limit.)<\/em><\/p>\n<p><!-- \/lmw:lab --><\/p>\n<p><!-- lmw:peers --><\/p>\n<h2>Recent Models in the Same Size Class<\/h2>\n<p><em>Models with <\/em><em>15\u201340B<\/em><em> parameters that Local Model Watch covered recently, listed by code from our article log for comparison. VRAM tiers are this site&#8217;s estimates; licenses are as stated on the model cards.<\/em><\/p>\n<div class=\"lmw-table-scroll\" tabindex=\"0\" style=\"overflow-x:auto;-webkit-overflow-scrolling:touch;max-width:100%;\">\n<p class=\"lmw-table-hint\" style=\"margin:0 0 4px;font-size:0.85em;opacity:0.7;\">\u2192 Scroll horizontally to see all columns<\/p>\n<table style=\"width:max-content;min-width:100%;border-collapse:collapse;\">\n<thead>\n<tr>\n<th>Model<\/th>\n<th>Parameters<\/th>\n<th>Smallest VRAM tier<\/th>\n<th>License<\/th>\n<th>Our article<\/th>\n<\/tr>\n<\/thead>\n<tbody>\n<tr>\n<td>elyza\/ELYZA-Thinking-1.0-llm-jp-4-32b-a3b<\/td>\n<td>32.1B<\/td>\n<td>80GB<\/td>\n<td>apache-2.0<\/td>\n<td><a href=\"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/2026\/10\/02\/elyza-thinking-1-0-32b-33b-2\/\">ELYZA Releases ELYZA-Thinking-1.0 32B\/33B Reasoning Models<\/a> (2026-10-02)<\/td>\n<\/tr>\n<tr>\n<td>XingChen-AGI\/Xing4.0-29B-A4B<\/td>\n<td>31.2B<\/td>\n<td>24GB<\/td>\n<td>apache-2.0<\/td>\n<td><a href=\"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/2026\/09\/28\/xing4-0-29b-a4b-china-telecom\/\">Xing4.0-29B-A4B Text Generation Model: Our Test Answers, 24GB+ VRAM<\/a> (2026-09-28)<\/td>\n<\/tr>\n<tr>\n<td>Edge0\/Edge0-35B-A3B-preview<\/td>\n<td>36.0B<\/td>\n<td>24GB<\/td>\n<td>apache-2.0<\/td>\n<td><a href=\"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/2026\/09\/11\/edge0-35b-a3b-preview-sparse-moe\/\">Edge0-35B-A3B-preview 35B MoE Model for Phone-Class Memory: 24GB+ VRAM<\/a> (2026-09-11)<\/td>\n<\/tr>\n<tr>\n<td>bartowski\/Gryphe_Pantheon-Reasoning-26B-A4B-1.1-V2-GGUF<\/td>\n<td>26.5B<\/td>\n<td>12GB<\/td>\n<td>apache-2.0<\/td>\n<td><a href=\"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/2026\/09\/11\/pantheon-reasoning-26b-gguf\/\">Gryphe_Pantheon-Reasoning-26B-A4B-1.1-V2-GGUF: Our Test Answers<\/a> (2026-09-11)<\/td>\n<\/tr>\n<tr>\n<td>nex-agi\/Nex-N2.5-mini<\/td>\n<td>35.1B<\/td>\n<td>16GB<\/td>\n<td>apache-2.0<\/td>\n<td><a href=\"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/2026\/09\/09\/nex-n25-mini-released\/\">Nex-N2.5-mini Text Generation Model: Our Test Answers, 16GB+ VRAM<\/a> (2026-09-09)<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<\/div>\n<p><!-- \/lmw:peers --><\/p>\n<h2>How to Get It<\/h2>\n<p>It is distributed in GGUF format and can be obtained from the Hugging Face repository <code>prism-ml\/Ternary-Bonsai-2-27B-gguf<\/code>. No license agreement (gated) is required.<\/p>\n<p>Because this model uses proprietary ternary hybrid attention kernels (<code>PTQ1_0<\/code> and <code>PQ2_0<\/code> packing), it will not work with standard llama.cpp. You must use the fork of llama.cpp provided by Prism ML or build it from source.<\/p>\n<p>The model can be downloaded using <code>huggingface-cli<\/code> (the <code>hf<\/code> command).<\/p>\n<pre><code class=\"language-bash\">hf download prism-ml\/Ternary-Bonsai-2-27B-gguf Ternary-Bonsai-2-27B-PQ2_0.gguf --local-dir.\n<\/code><\/pre>\n<p>An example execution using the Prism ML version of llama.cpp (<code>llama-cli<\/code>) is as follows:<\/p>\n<pre><code class=\"language-bash\">.\/bin\/llama-cli -m Ternary-Bonsai-2-27B-PQ2_0.gguf \\\n    -ngl 99 -fa on -c 32768 \\\n    --temp 1.0 --top-p 0.95 --top-k 20 \\\n    -p &quot;Explain quantum computing in simple terms.&quot; -n 256\n<\/code><\/pre>\n<p>Note that a separate repository for MLX (<code>Ternary-Bonsai-2-27B-mlx-2bit<\/code>) is also available for Apple Silicon.<\/p>\n<p><!-- lmw:variants --><\/p>\n<h2>Quantized and Converted Variants<\/h2>\n<div class=\"lmw-table-scroll\" tabindex=\"0\" style=\"overflow-x:auto;-webkit-overflow-scrolling:touch;max-width:100%;\">\n<p class=\"lmw-table-hint\" style=\"margin:0 0 4px;font-size:0.85em;opacity:0.7;\">\u2192 Scroll horizontally to see all columns<\/p>\n<table style=\"width:max-content;min-width:100%;border-collapse:collapse;\">\n<thead>\n<tr>\n<th>Added<\/th>\n<th>Publisher<\/th>\n<th>Format<\/th>\n<th>Repository<\/th>\n<th>Smallest VRAM tier (build, est. memory)<\/th>\n<\/tr>\n<\/thead>\n<tbody>\n<tr>\n<td>2026-09-18<\/td>\n<td>prism-ml<\/td>\n<td>GGUF<\/td>\n<td><a href=\"https:\/\/huggingface.co\/prism-ml\/Ternary-Bonsai-2-27B-gguf-dev\">prism-ml\/Ternary-Bonsai-2-27B-gguf-dev<\/a><\/td>\n<td>Q2_0 8.5GB (fits in 12GB VRAM)<\/td>\n<\/tr>\n<tr>\n<td>2026-09-24 (6 days after publication)<\/td>\n<td>prism-ml<\/td>\n<td>MLX<\/td>\n<td><a href=\"https:\/\/huggingface.co\/prism-ml\/Ternary-Bonsai-2-27B-mlx-2bit\">prism-ml\/Ternary-Bonsai-2-27B-mlx-2bit<\/a><\/td>\n<td>MLX 2bit 9.6GB (fits in 12GB VRAM)<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<\/div>\n<p>File sizes of each build:<\/p>\n<ul>\n<li>Available builds in prism-ml\/Ternary-Bonsai-2-27B-gguf-dev: Q2_0 7.1GB<\/li>\n<li>Available builds in prism-ml\/Ternary-Bonsai-2-27B-mlx-2bit: MLX 2bit 8.0GB<\/li>\n<\/ul>\n<p>In addition, 21 converted build(s) from other uploaders exist on Hugging Face; this site lists only builds from the model&#8217;s publisher or established quantization maintainers.<\/p>\n<p><em>This section is appended automatically by Local Model Watch when a converted build of this model appears after publication. Memory figures are estimated from the size of the distributed files. What the quantization names mean: <a href=\"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/glossary-quantization-en\/\">glossary<\/a>.<\/em><\/p>\n<p><!-- \/lmw:variants --><\/p>\n<p><!-- lmw:related --><\/p>\n<h2>Related Articles<\/h2>\n<ul>\n<li><a href=\"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/2026\/09\/26\/qwen3-8-27b-overview-specs-performance\/\">Qwen3.8-27B Vision-Language Model: Our Test Answers, 8GB+ VRAM<\/a><\/li>\n<li><a href=\"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/2026\/10\/02\/cloudflare-clef-27b-multimodal-decision-model\/\">clef Structured Decision-Making Model: 12GB+ VRAM, GGUF Builds<\/a><\/li>\n<li><a href=\"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/2026\/09\/28\/orcasaq2-27b-qwen3-8-27b-quantized\/\">OrcaSAQ-2-27B Text Generation Model: Our Test Answers, 16GB+ VRAM<\/a><\/li>\n<li><a href=\"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/2026\/09\/28\/hemmingway-1-open-27b-model-specialized-for-human-like-writing\/\">Hemmingway-1 Text Generation Model: 12GB+ VRAM, GGUF Builds<\/a><\/li>\n<\/ul>\n<p><!-- \/lmw:related --><\/p>\n<p><!-- lmw:next-steps --><\/p>\n<h2>What to Read Next<\/h2>\n<ul>\n<li><strong>Find models by VRAM<\/strong> (This model runs from the 8GB tier) \u2192 <a href=\"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/vram-guide-8gb-en\/\">Other models that run on a 8GB GPU<\/a><\/li>\n<li><strong>Explore the same model family<\/strong> \u2192 <a href=\"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/model-qwen-qwen3-8-en\/\">Qwen3.8 family overview (8 articles, 16 converted builds)<\/a><\/li>\n<li><strong>Engines that run this model<\/strong> \u2192 <a href=\"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/engine-llama-cpp-en\/\">llama.cpp<\/a> \/ <a href=\"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/engine-ollama-en\/\">Ollama<\/a><\/li>\n<li><strong>What Q1_0, Q2_0 mean and where to get this model<\/strong> \u2192 <a href=\"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/format-gguf-en\/\">GGUF format guide and models<\/a> \/ <a href=\"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/format-mlx-en\/\">MLX format guide and models<\/a> \/ <a href=\"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/glossary-quantization-en\/\">Quantization and model-format glossary<\/a><\/li>\n<li><strong>Learn about the publisher<\/strong> \u2192 <a href=\"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/publisher-alibaba-en\/\">Alibaba (Qwen): models, licenses and articles<\/a><\/li>\n<\/ul>\n<p><!-- \/lmw:next-steps --><\/p>\n<h2>Sources<\/h2>\n<ul>\n<li><a href=\"https:\/\/huggingface.co\/prism-ml\/Ternary-Bonsai-2-27B-gguf\">prism-ml\/Ternary-Bonsai-2-27B-gguf<\/a><\/li>\n<li><a href=\"https:\/\/huggingface.co\/Qwen\/Qwen3.8-27B\">Qwen\/Qwen3.8-27B<\/a><\/li>\n<li><a href=\"https:\/\/github.com\/PrismML-Eng\/Bonsai-demo\/blob\/main\/bonsai-2-27b-whitepaper.pdf\">Bonsai 2 27B Whitepaper<\/a><\/li>\n<li><a href=\"https:\/\/github.com\/PrismML-Eng\/Bonsai-demo\">Bonsai Demo &amp; Examples Repository<\/a><\/li>\n<li><a href=\"https:\/\/github.com\/PrismML-Eng\/llama.cpp\">llama.cpp (PrismML Fork)<\/a><\/li>\n<\/ul>\n<p><!-- lmw:updates --><\/p>\n<h2>Update History<\/h2>\n<ul>\n<li>2026-09-18: Added converted builds to \u201cQuantized and Converted Variants\u201d: prism-ml\/Ternary-Bonsai-2-27B-gguf-dev<\/li>\n<li>2026-09-24: Added converted builds to \u201cQuantized and Converted Variants\u201d: prism-ml\/Ternary-Bonsai-2-27B-mlx-2bit<\/li>\n<li>2026-09-24: Rebuilt the article: Added a &#8220;can you run it locally?&#8221; verdict from the distribution formats, converted builds, engine registries, license terms and measured compression<\/li>\n<li>2026-09-25: Changed the title to show what the article covers (VRAM requirements, file list, etc.).<\/li>\n<li>2026-09-28: Added our own measurements: Japanese token efficiency.<\/li>\n<li>2026-10-06: Added our own measurements: what is inside the GGUF file.<\/li>\n<li>2026-10-11: Added our own measurements: GPU speed, VRAM and answers to Japanese questions.<\/li>\n<li>2026-10-11: Changed the title to show what the article covers (VRAM requirements, file list, etc.).<\/li>\n<\/ul>\n<p><!-- \/lmw:updates --><\/p>\n","protected":false},"excerpt":{"rendered":"<p>Discover Ternary-Bonsai-2-27B-gguf, a ternary language model by Prism ML achieving 1.72 bits\/weight with near FP16 performance.<\/p>\n","protected":false},"author":1,"featured_media":11000,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[310],"tags":[1549,163,1551,520,522,1553,1547,1555],"class_list":["post-1776","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-new-models","tag-bonsai-2-en","tag-gguf-en","tag-prism-ml-en","tag-qwen-en","tag-qwen3-8-en","tag-ternary-en","tag-verified","tag--en"],"lang":"en","translations":{"en":1776,"ja":1774},"_links":{"self":[{"href":"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/wp-json\/wp\/v2\/posts\/1776","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/wp-json\/wp\/v2\/comments?post=1776"}],"version-history":[{"count":41,"href":"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/wp-json\/wp\/v2\/posts\/1776\/revisions"}],"predecessor-version":[{"id":11898,"href":"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/wp-json\/wp\/v2\/posts\/1776\/revisions\/11898"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/wp-json\/wp\/v2\/media\/11000"}],"wp:attachment":[{"href":"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/wp-json\/wp\/v2\/media?parent=1776"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/wp-json\/wp\/v2\/categories?post=1776"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/wp-json\/wp\/v2\/tags?post=1776"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}