{"id":9041,"date":"2026-10-02T12:09:38","date_gmt":"2026-10-02T03:09:38","guid":{"rendered":"https:\/\/localmodelwatch.tsuchitsuchi.com\/2026\/10\/02\/elyza-thinking-1-0-32b-33b-2\/"},"modified":"2026-11-17T11:08:31","modified_gmt":"2026-11-17T02:08:31","slug":"elyza-thinking-1-0-32b-33b-2","status":"publish","type":"post","link":"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/2026\/10\/02\/elyza-thinking-1-0-32b-33b-2\/","title":{"rendered":"ELYZA Releases ELYZA-Thinking-1.0 32B\/33B Reasoning Models"},"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\/elyza\/ELYZA-Thinking-1.0-llm-jp-4-32b-a3b\">elyza\/ELYZA-Thinking-1.0-llm-jp-4-32b-a3b<\/a><\/td>\n<\/tr>\n<tr>\n<td>Publisher guide<\/td>\n<td><a href=\"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/publisher-elyza-en\/\">ELYZA: models and licenses<\/a><\/td>\n<\/tr>\n<tr>\n<td>Published<\/td>\n<td>2026-10-02<\/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-safetensors-en\/\">safetensors<\/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 the publisher&#8217;s own code (transformers 5.18.0, bfloat16, revision <code>5260ecc249<\/code>) on a cloud GPU on Modal (NVIDIA A100-SXM4-80GB, peak VRAM 60.1GB) and show its answers verbatim.<\/li>\n<li>It needs about 40% fewer tokens than the Qwen3 tokenizer for the same Japanese text, so about 1.66\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>ELYZA, Inc. has released ELYZA-Thinking-1.0, a reasoning model in 32B and 33B sizes based on the llm-jp-4 series of fully domestic Japanese foundational models. Among the released models, the 32B model (elyza\/ELYZA-Thinking-1.0-llm-jp-4-32b-a3b) adopts an MoE architecture and has undergone mid-training and post-training to improve Japanese and English language understanding and generation capabilities. It specializes in Japanese-specific knowledge retrieval, instruction following with complex constraints, and reasoning capabilities in mathematics, coding, and STEM fields.<\/p>\n<h2>Specifications<\/h2>\n<ul>\n<li>Parameters: 32.1B (32B-A3B) \/ 33B (33B model)<\/li>\n<li>Architecture: Qwen3MoeForCausalLM (32B) \/ LlamaForCausalLM (33B)<\/li>\n<li>Context Length: 65,536<\/li>\n<li>Active Parameters: 3,827,476,992 (32B)<\/li>\n<li>Routed Experts: 128 (32B)<\/li>\n<li>Active Experts: 8 (32B)<\/li>\n<li>Hidden Size: 2,560 (32B) \/ 5,120 (33B)<\/li>\n<li>Layers: 32 (32B) \/ 64 (33B)<\/li>\n<li>Heads: 40 (32B) \/ 40 (33B)<\/li>\n<\/ul>\n<h2>Performance<\/h2>\n<p>The benchmark comparison table for the MoE model (ELYZA-thinking-1.0-llm-jp-4-32b-a3b) as listed in the model card is as follows. Scores for same-family models and other open-weight models are shown for comparison.<\/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>Benchmark<\/th>\n<th>ELYZA-thinking-1.0- llm-jp-4- 32b-a3b<\/th>\n<th>llm-jp-4- 32b-a3b- thinking<\/th>\n<th>llm-jp-4.1- 32b-a3b- thinking<\/th>\n<th>Qwen3- 30B-A3B<\/th>\n<th>gpt-oss-20b<\/th>\n<th>Nemotron 3.5 Lightning 30B-A3B<\/th>\n<th>Qwen3.5- 35B-A3B<\/th>\n<th>Gemma 4 26B-A4B<\/th>\n<\/tr>\n<\/thead>\n<tbody>\n<tr>\n<td>Knowledge &amp; STEM<\/td>\n<td><\/td>\n<td><\/td>\n<td><\/td>\n<td><\/td>\n<td><\/td>\n<td><\/td>\n<td><\/td>\n<td><\/td>\n<\/tr>\n<tr>\n<td>MMLU-Pro (en)<\/td>\n<td>73.14<\/td>\n<td>69.70<\/td>\n<td>73.90<\/td>\n<td>78.29<\/td>\n<td>75.50<\/td>\n<td>79.70<\/td>\n<td>84.12<\/td>\n<td>82.90<\/td>\n<\/tr>\n<tr>\n<td>JMMLU (ja)<\/td>\n<td>81.05<\/td>\n<td>81.32<\/td>\n<td>83.09<\/td>\n<td>84.25<\/td>\n<td>82.79<\/td>\n<td>85.88<\/td>\n<td>89.76<\/td>\n<td>88.33<\/td>\n<\/tr>\n<tr>\n<td>GPQA-Diamond (en)<\/td>\n<td>63.86<\/td>\n<td>51.93<\/td>\n<td>60.80<\/td>\n<td>63.26<\/td>\n<td>68.59<\/td>\n<td>75.98<\/td>\n<td>84.38<\/td>\n<td>78.60<\/td>\n<\/tr>\n<tr>\n<td>GPQA (ja)<\/td>\n<td>57.24<\/td>\n<td>51.03<\/td>\n<td>57.06<\/td>\n<td>56.57<\/td>\n<td>63.80<\/td>\n<td>67.30<\/td>\n<td>77.08<\/td>\n<td>74.82<\/td>\n<\/tr>\n<tr>\n<td>Math<\/td>\n<td><\/td>\n<td><\/td>\n<td><\/td>\n<td><\/td>\n<td><\/td>\n<td><\/td>\n<td><\/td>\n<td><\/td>\n<\/tr>\n<tr>\n<td>MATH-500 (en)<\/td>\n<td>95.20<\/td>\n<td>84.20<\/td>\n<td>93.80<\/td>\n<td>97.20<\/td>\n<td>97.40<\/td>\n<td>98.00<\/td>\n<td>99.00<\/td>\n<td>98.80<\/td>\n<\/tr>\n<tr>\n<td>JMATH-500 (ja)<\/td>\n<td>88.20<\/td>\n<td>83.20<\/td>\n<td>87.40<\/td>\n<td>91.60<\/td>\n<td>92.20<\/td>\n<td>94.00<\/td>\n<td>93.20<\/td>\n<td>92.00<\/td>\n<\/tr>\n<tr>\n<td>AIME 2024+2025 (en)<\/td>\n<td>62.29<\/td>\n<td>34.48<\/td>\n<td>57.71<\/td>\n<td>76.25<\/td>\n<td>88.23<\/td>\n<td>89.90<\/td>\n<td>92.92<\/td>\n<td>90.62<\/td>\n<\/tr>\n<tr>\n<td>PolyMath (ja, high+top)<\/td>\n<td>34.60<\/td>\n<td>12.80<\/td>\n<td>27.40<\/td>\n<td>43.45<\/td>\n<td>55.10<\/td>\n<td>58.20<\/td>\n<td>56.40<\/td>\n<td>65.10<\/td>\n<\/tr>\n<tr>\n<td>Japanese QA<\/td>\n<td><\/td>\n<td><\/td>\n<td><\/td>\n<td><\/td>\n<td><\/td>\n<td><\/td>\n<td><\/td>\n<td><\/td>\n<\/tr>\n<tr>\n<td>JamC-QA (ja)<\/td>\n<td>59.51<\/td>\n<td>53.75<\/td>\n<td>53.79<\/td>\n<td>45.30<\/td>\n<td>41.10<\/td>\n<td>55.17<\/td>\n<td>60.29<\/td>\n<td>66.18<\/td>\n<\/tr>\n<tr>\n<td>JEMHopQA (ja)<\/td>\n<td>71.59<\/td>\n<td>62.71<\/td>\n<td>65.51<\/td>\n<td>55.20<\/td>\n<td>54.16<\/td>\n<td>55.17<\/td>\n<td>58.88<\/td>\n<td>61.93<\/td>\n<\/tr>\n<tr>\n<td>Instruction Following<\/td>\n<td><\/td>\n<td><\/td>\n<td><\/td>\n<td><\/td>\n<td><\/td>\n<td><\/td>\n<td><\/td>\n<td><\/td>\n<\/tr>\n<tr>\n<td>IFEval (en)<\/td>\n<td>93.70<\/td>\n<td>83.12<\/td>\n<td>92.75<\/td>\n<td>89.21<\/td>\n<td>88.85<\/td>\n<td>94.96<\/td>\n<td>88.13<\/td>\n<td>95.35<\/td>\n<\/tr>\n<tr>\n<td>IFBench (en)<\/td>\n<td>59.16<\/td>\n<td>49.64<\/td>\n<td>56.18<\/td>\n<td>36.55<\/td>\n<td>60.10<\/td>\n<td>71.95<\/td>\n<td>59.96<\/td>\n<td>71.95<\/td>\n<\/tr>\n<tr>\n<td>M-IFEval-ja (ja)<\/td>\n<td>81.75<\/td>\n<td>63.05<\/td>\n<td>77.32<\/td>\n<td>63.72<\/td>\n<td>73.12<\/td>\n<td>79.76<\/td>\n<td>76.88<\/td>\n<td>87.72<\/td>\n<\/tr>\n<tr>\n<td>JFBench (ja)<\/td>\n<td>38.15<\/td>\n<td>27.14<\/td>\n<td>30.87<\/td>\n<td>24.23<\/td>\n<td>18.38<\/td>\n<td>34.39<\/td>\n<td>28.83<\/td>\n<td>30.30<\/td>\n<\/tr>\n<tr>\n<td>Translation<\/td>\n<td><\/td>\n<td><\/td>\n<td><\/td>\n<td><\/td>\n<td><\/td>\n<td><\/td>\n<td><\/td>\n<td><\/td>\n<\/tr>\n<tr>\n<td>WMT20 en\u2192ja (ja)<\/td>\n<td>23.98<\/td>\n<td>23.53<\/td>\n<td>23.76<\/td>\n<td>22.17<\/td>\n<td>23.20<\/td>\n<td>20.29<\/td>\n<td>24.89<\/td>\n<td>27.90<\/td>\n<\/tr>\n<tr>\n<td>WMT20 ja\u2192en (en)<\/td>\n<td>22.00<\/td>\n<td>20.28<\/td>\n<td>20.41<\/td>\n<td>21.58<\/td>\n<td>21.35<\/td>\n<td>20.34<\/td>\n<td>23.70<\/td>\n<td>26.22<\/td>\n<\/tr>\n<tr>\n<td>Coding<\/td>\n<td><\/td>\n<td><\/td>\n<td><\/td>\n<td><\/td>\n<td><\/td>\n<td><\/td>\n<td><\/td>\n<td><\/td>\n<\/tr>\n<tr>\n<td>LiveCodeBench v6 (en)<\/td>\n<td>53.66<\/td>\n<td>26.86<\/td>\n<td>32.40<\/td>\n<td>56.40<\/td>\n<td>59.94<\/td>\n<td>75.94<\/td>\n<td>74.91<\/td>\n<td>77.09<\/td>\n<\/tr>\n<tr>\n<td>JHumanEval (ja)<\/td>\n<td>95.37<\/td>\n<td>91.52<\/td>\n<td>92.38<\/td>\n<td>93.72<\/td>\n<td>74.63<\/td>\n<td>96.83<\/td>\n<td>93.54<\/td>\n<td>98.35<\/td>\n<\/tr>\n<tr>\n<td>Function Calling<\/td>\n<td><\/td>\n<td><\/td>\n<td><\/td>\n<td><\/td>\n<td><\/td>\n<td><\/td>\n<td><\/td>\n<td><\/td>\n<\/tr>\n<tr>\n<td>BFCL v4 (en)<\/td>\n<td>36.31<\/td>\n<td>13.29<\/td>\n<td>27.96<\/td>\n<td>42.65<\/td>\n<td>48.09<\/td>\n<td>62.37<\/td>\n<td>67.03<\/td>\n<td>66.88<\/td>\n<\/tr>\n<tr>\n<td>Nejumi BFCL (ja)<\/td>\n<td>41.89<\/td>\n<td>11.58<\/td>\n<td>39.19<\/td>\n<td>49.03<\/td>\n<td>47.10<\/td>\n<td>52.12<\/td>\n<td>60.62<\/td>\n<td>57.34<\/td>\n<\/tr>\n<tr>\n<td>Average<\/td>\n<td><\/td>\n<td><\/td>\n<td><\/td>\n<td><\/td>\n<td><\/td>\n<td><\/td>\n<td><\/td>\n<td><\/td>\n<\/tr>\n<tr>\n<td>Overall<\/td>\n<td>58.46<\/td>\n<td>46.38<\/td>\n<td>53.89<\/td>\n<td>56.31<\/td>\n<td>57.26<\/td>\n<td>64.52<\/td>\n<td>66.37<\/td>\n<td>68.58<\/td>\n<\/tr>\n<tr>\n<td>Japanese<\/td>\n<td>59.61<\/td>\n<td>49.16<\/td>\n<td>56.65<\/td>\n<td>56.73<\/td>\n<td>55.04<\/td>\n<td>62.02<\/td>\n<td>64.24<\/td>\n<td>66.68<\/td>\n<\/tr>\n<tr>\n<td>English<\/td>\n<td>55.94<\/td>\n<td>41.16<\/td>\n<td>49.72<\/td>\n<td>56.84<\/td>\n<td>61.45<\/td>\n<td>68.98<\/td>\n<td>69.98<\/td>\n<td>71.55<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<\/div>\n<p>According to evaluations by the publishers, ELYZA-Thinking-1.0-llm-jp-4-32b-a3b demonstrates high performance on Japanese-related tasks. In particular, it records excellent figures in IFEval and M-IFEval-ja, which measure instruction-following capability, as well as in Japanese coding via JHumanEval, showing overall score improvements compared to the base model group. On the other hand, in Function Calling and advanced English mathematics\/competitive programming benchmarks (such as AIME and LiveCodeBench v6), it scores lower than some overseas large-scale models and other models, indicating room for improvement in specific English-language tasks and tool selection accuracy.<\/p>\n<h2>Strengths and Use Cases<\/h2>\n<p>The ELYZA-Thinking-1.0 series is built through a three-stage training process consisting of mid-training, SFT (supervised fine-tuning), and reinforcement learning (RLVR) to enhance Japanese and English language understanding and generation capabilities. It has been trained on localized Japanese reasoning data for mathematics, coding, and STEM fields, specializing in Japanese instruction following and knowledge retrieval. Additionally, it leverages synthetic knowledge data from Japanese Wikipedia and Wikidata to strengthen the understanding and retrieval of Japanese-specific knowledge. Furthermore, it supports tool calling and agent use cases, with agent data localized into Japanese while maintaining tool calls, function names, arguments, and JSON structures.<\/p>\n<p><!-- lmw:hardware --><\/p>\n<h2>Hardware Requirements<\/h2>\n<p><strong>Estimated requirements (calculated by Local Model Watch)<\/strong> \u2014 32.1B parameters<\/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>80GB class (A100 \/ H100)<\/td>\n<td>BF16<\/td>\n<td>59.9GB<\/td>\n<td>71.8GB<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<\/div>\n<p><strong>Inference engine support<\/strong> (architecture name matched against each project&#8217;s own model registry in its source code, checked 2026-11-17): llama.cpp: registered, vLLM: registered, MLX (mlx-lm): registered.<\/p>\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>Not usable in Ollama, LM Studio and llama.cpp yet \u2014 we have found no GGUF build.<\/strong><\/p>\n<p>The publisher ships safetensors only. However, llama.cpp&#8217;s registry does list this architecture, so <strong>conversion to GGUF is possible<\/strong> and the model will run once someone publishes a converted build. 3 converted build(s) from other uploaders exist. Today it can be run with transformers or vLLM, using the memory figures in the table above.<\/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 we cannot run this model on 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>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>414<\/strong><\/td>\n<td><strong>0.74\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 40% fewer tokens than the Qwen3 tokenizer for the same Japanese text, so about 1.66\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\/elyza\/ELYZA-Thinking-1.0-llm-jp-4-32b-a3b\">elyza\/ELYZA-Thinking-1.0-llm-jp-4-32b-a3b<\/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>Running It With the Publisher&#8217;s Own Code (Cloud GPU)<\/h3>\n<p>We have not found a GGUF version of this model, and the conversion script of llama.cpp b11541 could not convert its tokenizer (the tokenizer settings require the publisher&#8217;s own code) when we tried on 2026-10-11, so we ran the original weights (BF16) with the Python code the publisher ships on Hugging Face (transformers with trust_remote_code). That code ran only inside a cloud container with its network cut off.<\/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>Result<\/th>\n<\/tr>\n<\/thead>\n<tbody>\n<tr>\n<td>VRAM right after loading the weights (BF16)<\/td>\n<td>59.9GB<\/td>\n<\/tr>\n<tr>\n<td>Peak VRAM while answering the five questions<\/td>\n<td>60.1GB<\/td>\n<\/tr>\n<tr>\n<td>Generation (median over the five answers, transformers)<\/td>\n<td>13.7 tokens\/s<\/td>\n<\/tr>\n<tr>\n<td>Generation in Japanese characters<\/td>\n<td>about 33.0 characters\/s<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<\/div>\n<p>Conditions: NVIDIA A100-SXM4-80GB (VRAM 79GB) on Modal, transformers 5.18.0, PyTorch 2.14.1+cu130, <a href=\"https:\/\/huggingface.co\/elyza\/ELYZA-Thinking-1.0-llm-jp-4-32b-a3b\/tree\/5260ecc249f32ef5005ef940c78f41f130d9c468\">elyza\/ELYZA-Thinking-1.0-llm-jp-4-32b-a3b<\/a> at revision <code>5260ecc249<\/code>, BF16, measured 2026-10-11 (JST).<\/p>\n<p>VRAM is what PyTorch allocated for the unquantized BF16 weights; a quantized GGUF would need far less. transformers generates much more slowly than engines such as llama.cpp, so treat the speed as a reference only.<\/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\">\u307f\u3069\u308a\u5e02\u306f\u6625\u3001\u6700\u53e4\u30fb\u4e2d\u592e\u56f3\u66f8\u9928\u3092\u7d041\u5e74\u534a\u30fb12\u5104\u5186\u3067\u6539\u4fee\u3057\u300110\u67081\u65e5\u306b\u518d\u958b\u9928\u3057\u305f\u3002  \n\u5ea7\u5e2d\u914d\u7f6e\u3092\u898b\u76f4\u3057\u3001\u5b66\u7fd2\u5e2d\u30923\u968e\u306b\u96c6\u7d04\u30011\u968e\u306f\u4f1a\u8a71\u53ef\u80fd\u306a\u300c\u306b\u304e\u308f\u3044\u30d5\u30ed\u30a2\u300d\u30012\u968e\u306f\u9759\u304b\u306a\u300c\u95b2\u89a7\u30d5\u30ed\u30a2\u300d\u306b\u3057\u3001\u30ab\u30d5\u30a7\u3084\u30b0\u30eb\u30fc\u30d7\u5b66\u7fd2\u5ba4\u3001\u96fb\u5b50\u66f8\u7c4d\u30b5\u30fc\u30d3\u30b9\u306a\u3069\u3092\u65b0\u8a2d\u3057\u305f\u3002  \n\u8a2d\u5099\u66f4\u65b0\u3084\u30d0\u30ea\u30a2\u30d5\u30ea\u30fc\u5316\u3067\u5229\u7528\u8005\u6e80\u8db3\u5ea6\u306f\u4e0a\u304c\u3063\u305f\u304c\u3001\u99d0\u8eca\u5834\u4e0d\u8db3\u3084\u8077\u54e1\u8ca0\u62c5\u5897\u3068\u3044\u3063\u305f\u8ab2\u984c\u304c\u6b8b\u308a\u3001\u4eca\u5f8c\u3082\u4ed6\u9928\u6539\u4fee\u306e\u691c\u8a0e\u304c\u7d9a\u304f\u3002\n<\/code><\/pre>\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\">\u4ef6\u540d\uff1a\u6253\u3061\u5408\u308f\u305b\u65e5\u7a0b\u5909\u66f4\u306e\u304a\u9858\u3044\uff08\u25cb\u6708\u25cb\u65e5\uff09\n\n\u25cb\u25cb\u682a\u5f0f\u4f1a\u793e  \n\u25b3\u25b3\u90e8\u3000\u25a1\u25a1\u69d8\n\n\u3044\u3064\u3082\u5927\u5909\u304a\u4e16\u8a71\u306b\u306a\u3063\u3066\u304a\u308a\u307e\u3059\u3002  \n\u682a\u5f0f\u4f1a\u793e\u25cb\u25cb\u306e\u25b3\u25b3\uff08\u81ea\u5206\u306e\u540d\u524d\uff09\u3067\u3059\u3002\n\n\u3055\u3066\u3001\u6765\u9031\u706b\u66dc\u65e5\uff08\u25cb\u6708\u25cb\u65e5\uff09\u306b\u4e88\u5b9a\u3057\u3066\u304a\u308a\u307e\u3059\u6253\u3061\u5408\u308f\u305b\u306b\u3064\u304d\u307e\u3057\u3066\u3001  \n\u8aa0\u306b\u6050\u7e2e\u3067\u3059\u304c\u3001\u540c\u65e5\u306e\u3054\u90fd\u5408\u304c\u3064\u304b\u306a\u304f\u306a\u3063\u305f\u305f\u3081\u3001  \n**\u6c34\u66dc\u65e5\uff08\u25cb\u6708\u25cb\u65e5\uff09\u306b\u5909\u66f4\u3055\u305b\u3066\u3044\u305f\u3060\u304f\u3053\u3068\u306f\u53ef\u80fd\u3067\u3057\u3087\u3046\u304b**\u3002\n\n\u3082\u3057\u6c34\u66dc\u65e5\u3082\u3054\u90fd\u5408\u304c\u60aa\u3044\u5834\u5408\u306f\u3001\u5225\u9014\u3054\u8abf\u6574\u3044\u305f\u3060\u3051\u308b\u65e5\u6642\u3092\u3054\u63d0\u793a\u3044\u305f\u3060\u3051\u307e\u3059\u3068\u5e78\u3044\u3067\u3059\u3002  \n\u3054\u591a\u5fd9\u306e\u3068\u3053\u308d\u6050\u308c\u5165\u308a\u307e\u3059\u304c\u3001\u4f55\u5352\u3054\u691c\u8a0e\u306e\u307b\u3069\u3088\u308d\u3057\u304f\u304a\u9858\u3044\u7533\u3057\u4e0a\u3052\u307e\u3059\u3002\n\n---\n\n**\u682a\u5f0f\u4f1a\u793e\u25cb\u25cb**  \n\u25b3\u25b3\u90e8\u3000\u25b3\u25b3\uff08\u81ea\u5206\u306e\u540d\u524d\uff09  \n\u96fb\u8a71\uff1a00-0000-0000  \n\u30e1\u30fc\u30eb\uff1aexample@company.co.jp  \n\n---  \n\n\u3054\u8fd4\u4fe1\u3092\u304a\u5f85\u3061\u3057\u3066\u304a\u308a\u307e\u3059\u3002  \n\n\u3088\u308d\u3057\u304f\u304a\u9858\u3044\u3044\u305f\u3057\u307e\u3059\u3002\n<\/code><\/pre>\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\">**\u8a08\u7b97\u624b\u9806**\n\n1. **\u308a\u3093\u3054\u306e\u4ee3\u91d1**  \n   1\u500b 120\u5186 \u00d7 3\u500b \uff1d 120\u5186 \u00d7 3 \uff1d **360\u5186**\n\n2. **\u307f\u304b\u3093\u306e\u4ee3\u91d1**  \n   1\u500b 45\u5186 \u00d7 8\u500b \uff1d 45\u5186 \u00d7 8 \uff1d **360\u5186**\n\n3. **\u5408\u8a08\u91d1\u984d**  \n   \u308a\u3093\u3054 360\u5186 \uff0b \u307f\u304b\u3093 360\u5186 \uff1d **720\u5186**\n\n4. **\u652f\u6255\u3063\u305f\u91d1\u984d**  \n   1,000\u5186\u672d\u3067\u652f\u6255\u3063\u305f\u306e\u3067 **1,000\u5186**\n\n5. **\u304a\u3064\u308a**  \n   1,000\u5186 \uff0d 720\u5186 \uff1d **280\u5186**\n\n---\n\n**\u7b54\u3048**  \n\u304a\u3064\u308a\u306f **280\u5186** \u3067\u3059\u3002\n<\/code><\/pre>\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<p><em>(The model spent the whole token budget thinking and gave no answer.)<\/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<p><em>(The model spent the whole token budget thinking and gave no answer.)<\/em><\/p>\n<h4>Setup, Download and Run Steps (What We Actually Ran)<\/h4>\n<p>We split the work into two containers on Modal, both built from the Debian (slim) image with Python 3.12. The container that downloads the files never runs the publisher&#8217;s code; the container that runs it has no network access. These are the packages we installed (without version pins; the versions actually installed are listed below).<\/p>\n<pre><code class=\"language-bash\"># 1. Environment for downloading (no GPU)\npython -m pip install huggingface_hub\n\n# 2. Environment for running (GPU, network blocked)\npython -m pip install torch transformers accelerate safetensors sentencepiece tiktoken protobuf einops compressed-tensors\n<\/code><\/pre>\n<p>We then called <code>download(job)<\/code> in the first container and <code>run(job)<\/code> in the second one from the script below (<code>remote_code_remote.py<\/code>, written by us; full text). Files go under <code>\/models\/&lt;dir&gt;<\/code>. Comments are in Japanese as in our repository:<\/p>\n<pre><code class=\"language-python\">&quot;&quot;&quot;\nModal \u306e\u30b3\u30f3\u30c6\u30ca\u306e\u4e2d\u3067\u52d5\u304f\u3001\u914d\u5e03\u5143\u306e\u30b3\u30fc\u30c9(trust_remote_code)\u306b\u3088\u308b\u5b9f\u6e2c(2026-09-30)\u3002\nlmw\/lab\/gpu_modal.py \u306e download_weights \/ run_remote_code \u304b\u3089\u547c\u3070\u308c\u308b\u3002**\u6a19\u6e96\u30e9\u30a4\u30d6\u30e9\u30ea\u3060\u3051\u3092\u5148\u982d\u3067\u8aad\u307f**\u3001\nhuggingface_hub\u30fbtorch\u30fbtransformers \u306f\u5404\u95a2\u6570\u306e\u4e2d\u3067\u8aad\u3080(\u30b3\u30f3\u30c6\u30ca\u3054\u3068\u306b\u5165\u3063\u3066\u3044\u308b\u3082\u306e\u304c\u9055\u3046)\u3002\n\n\u30102\u3064\u306e\u30b3\u30f3\u30c6\u30ca\u306b\u5206\u3051\u308b\u3011\n- download: GPU \u306e\u7121\u3044\u30b3\u30f3\u30c6\u30ca\u30fb\u30cd\u30c3\u30c8\u30ef\u30fc\u30af\u3042\u308a\u3002\u30d5\u30a1\u30a4\u30eb\u3092\u843d\u3068\u3059\u3060\u3051\u3067\u3001**\u914d\u5e03\u5143\u306e\u30b3\u30fc\u30c9\u306f\u5b9f\u884c\u3057\u306a\u3044**\n  (import \u3082\u3057\u306a\u3044)\u3002\u91cd\u307f\u306f safetensors \u3060\u3051(\u547c\u3073\u51fa\u3057\u5074\u304c\u6e21\u3059\u30d5\u30a1\u30a4\u30eb\u540d\u306e\u4e00\u89a7\u306b pickle \u306f\u5165\u3089\u306a\u3044)\u3002\n- run: GPU \u306e\u30b3\u30f3\u30c6\u30ca\u30fb**\u30cd\u30c3\u30c8\u30ef\u30fc\u30af\u906e\u65ad**(block_network)\u30fbModal \u306e\u6a5f\u80fd\u3082\u4f7f\u3048\u306a\u3044(restrict_modal_access)\u30fb\n  Volume \u306f\u8aad\u307f\u53d6\u308a\u5c02\u7528\u3002\u3053\u3053\u3067\u3060\u3051\u914d\u5e03\u5143\u306e\u30b3\u30fc\u30c9\u3092\u5b9f\u884c\u3059\u308b\u3002\u5916\u3078\u9001\u308b\u3053\u3068\u3082\u3001\u307b\u304b\u306e\u30b3\u30f3\u30c6\u30ca\u3092\u8d77\u52d5\u3059\u308b\u3053\u3068\u3082\u3067\u304d\u306a\u3044\u3002\n\n\u8a18\u4e8b\u57f7\u7b46\u7528\u306e LLM \u306f\u547c\u3070\u306a\u3044\u3002\u56de\u7b54\u306e\u826f\u3057\u60aa\u3057\u3082\u5224\u5b9a\u3057\u306a\u3044(\u539f\u6587\u3092\u8fd4\u3059\u3060\u3051)\u3002\n&quot;&quot;&quot;\nimport json\nimport os\nimport re\nimport platform\nimport time\nfrom datetime import datetime, timezone\n\nMOUNT = &quot;\/models&quot;\n# \u30b3\u30f3\u30c6\u30ca\u306e\u30a4\u30e1\u30fc\u30b8(lmw\/lab\/gpu_modal.py \u304c\u4f7f\u3044\u3001\u8a18\u4e8b\u306e\u624b\u9806\u306b\u3082\u540c\u3058\u5024\u3092\u8f09\u305b\u308b)\nPYTHON_VERSION = &quot;3.12&quot;\nDOWNLOAD_PACKAGES = (&quot;huggingface_hub&quot;,)\n# compressed-tensors \u306f\u91cf\u5b50\u5316\u6e08\u307f\u306e\u914d\u5e03(NVFP4 \u7b49\u30022026-10-01\u3001K2-Horizon-32B-NVFP4)\u3092 transformers \u304c\u8aad\u3080\u305f\u3081\u306b\u5165\u308c\u308b\nRUN_PACKAGES = (&quot;torch&quot;, &quot;transformers&quot;, &quot;accelerate&quot;, &quot;safetensors&quot;, &quot;sentencepiece&quot;, &quot;tiktoken&quot;, &quot;protobuf&quot;, &quot;einops&quot;,\n                &quot;compressed-tensors&quot;)\n\ndef installed_packages() -&gt; list[str]:\n    &quot;&quot;&quot;\u3053\u306e\u30b3\u30f3\u30c6\u30ca\u306b\u5165\u3063\u3066\u3044\u308b\u30d1\u30c3\u30b1\u30fc\u30b8\u306e\u7248(pip freeze \u3068\u540c\u3058\u5f62)\u3002\u8a18\u4e8b\u306e\u624b\u9806\u306b\u8f09\u305b\u308b\u3002&quot;&quot;&quot;\n    from importlib import metadata\n    seen = {}\n    for dist in metadata.distributions():\n        name = dist.metadata[&quot;Name&quot;]\n        if name and name.lower() not in seen:\n            seen[name.lower()] = f&quot;{name}=={dist.version}&quot;\n    return sorted(seen.values(), key=str.lower)\n\ndef _log(t0: float, message: str) -&gt; None:\n    print(f&quot;[rc {time.time() - t0:6.0f}s] {message}&quot;, flush=True)\n\ndef download(job: dict) -&gt; dict:\n    &quot;&quot;&quot;job: {&quot;repo&quot;, &quot;revision&quot;, &quot;files&quot;: [...], &quot;dir&quot;}\u3002\u623b\u308a\u5024: {&quot;bytes&quot;, &quot;seconds&quot;, &quot;packages&quot;}\u3002&quot;&quot;&quot;\n    from huggingface_hub import hf_hub_download\n    t0 = time.time()\n    target = os.path.join(MOUNT, job[&quot;dir&quot;])\n    os.makedirs(target, exist_ok=True)\n    total = 0\n    for name in job[&quot;files&quot;]:\n        path = hf_hub_download(job[&quot;repo&quot;], name, revision=job[&quot;revision&quot;], local_dir=target)\n        total += os.path.getsize(path)\n    seconds = round(time.time() - t0)\n    _log(t0, f&quot;\u30c0\u30a6\u30f3\u30ed\u30fc\u30c9 {total \/ 1024 ** 3:.1f}GB\u30fb{seconds}\u79d2({total \/ 1024 ** 2 \/ max(seconds, 1):.0f}MB\/\u79d2)&quot;)\n    return {&quot;bytes&quot;: int(total), &quot;seconds&quot;: int(seconds), &quot;packages&quot;: installed_packages()}\n\ndef _chat_ids(tok, text: str):\n    messages = [{&quot;role&quot;: &quot;user&quot;, &quot;content&quot;: text}]\n    try:\n        out = tok.apply_chat_template(messages, add_generation_prompt=True, enable_thinking=False,\n                                      return_tensors=&quot;pt&quot;)\n    except TypeError:\n        out = tok.apply_chat_template(messages, add_generation_prompt=True, return_tensors=&quot;pt&quot;)\n    # \u7248\u306b\u3088\u3063\u3066\u306f dict(BatchEncoding)\u304c\u8fd4\u308b\n    return out[&quot;input_ids&quot;] if hasattr(out, &quot;keys&quot;) else out\n\n# \u601d\u8003\u306e\u76ee\u5370\u3002&lt;think&gt; \u306b\u9650\u3089\u306a\u3044(K2-Horizon \u306f &lt;ifm|think&gt;\u30fb&lt;ifm|think_fast&gt; \u3068 &lt;\/ifm|think&gt;\u30022026-09-30)\n_THINK_OPEN_RE = re.compile(r&quot;&lt;(?!\/)([^&lt;&gt;\\s]*think[^&lt;&gt;\\s]*)&gt;&quot;)\n\n# Harmony \u5f62\u5f0f(gpt-oss\u30fbLLM-jp-4)\u306e1\u3064\u306e\u30e1\u30c3\u30bb\u30fc\u30b8: &lt;|channel|&gt;\u30c1\u30e3\u30f3\u30cd\u30eb\u540d&lt;|message|&gt;\u672c\u6587\u3002\u672c\u6587\u306f\u6b21\u306e\u8a18\u53f7\u306e\u524d\u307e\u3067\u3002\n# LLM-jp-4 \u306f\u8a18\u53f7\u306e\u5f8c\u306b\u7a7a\u767d\u304c\u5165\u308b(llama.cpp \u306e common\/parsers\/llm-jp-harmony.cpp)\n_HARMONY_RE = re.compile(r&quot;&lt;\\|channel\\|&gt;\\s*([A-Za-z]+)[^&lt;]*?&lt;\\|message\\|&gt;(.*?)(?=&lt;\\|end\\|&gt;|&lt;\\|return\\|&gt;|&lt;\\|call\\|&gt;|&lt;\\|start\\|&gt;|$)&quot;,\n                         re.S)\n\ndef split_thinking(generated: str, prompt_text: str = &quot;&quot;) -&gt; tuple[str, str]:\n    &quot;&quot;&quot;\n    \u751f\u6210\u3057\u305f\u6587\u5b57\u5217(\u7279\u5225\u306a\u8a18\u53f7\u3092\u6b8b\u3057\u305f\u3082\u306e)\u3092 (\u56de\u7b54, \u601d\u8003) \u306b\u5206\u3051\u308b\u3002llama-server \u306e reasoning_content \u3068\u540c\u3058\u6271\u3044\u3002\n    \u30c1\u30e3\u30c3\u30c8\u30c6\u30f3\u30d7\u30ec\u30fc\u30c8\u304c\u30d7\u30ed\u30f3\u30d7\u30c8\u306e\u672b\u5c3e\u3067\u601d\u8003\u306e\u76ee\u5370\u3092\u958b\u3051\u3066\u3044\u308b\u5834\u5408\u306f\u3001\u751f\u6210\u306f\u601d\u8003\u306e\u9014\u4e2d\u304b\u3089\u59cb\u307e\u308b\u3002\n    \u9589\u3058\u308b\u76ee\u5370\u306f\u3001\u958b\u3051\u305f\u76ee\u5370\u306e\u540d\u524d\u304b\u3089\u6c7a\u3081\u308b(&lt;X&gt; \u2192 &lt;\/X&gt;\u3002\u305f\u3060\u3057 think_fast \u7b49\u306e\u5909\u7a2e\u306f &lt;\/\u2026think&gt; \u3067\u3082\u9589\u3058\u308b)\u3002\n    2026-09-30\u3001\u7279\u5225\u306a\u8a18\u53f7\u3054\u3068\u6d88\u3057\u3066\u6587\u5b57\u306b\u3057\u3001\u6b21\u306b &lt;think&gt; \u3060\u3051\u3092\u524d\u63d0\u306b\u3057\u305f\u305f\u3081\u3001K2-Horizon \u306e\u82f1\u8a9e\u306e\u601d\u8003\u304c\n    2\u5ea6\u300c\u56de\u7b54\u300d\u3068\u3057\u3066\u8a18\u9332\u3055\u308c\u305f\u3002\n    Harmony \u5f62\u5f0f(&lt;|channel|&gt;analysis \u304c\u601d\u8003\u3001&lt;|channel|&gt;final \u304c\u56de\u7b54)\u306f\u30c1\u30e3\u30f3\u30cd\u30eb\u3067\u5206\u3051\u308b(2026-10-03\u3001\n    ELYZA-Thinking-1.0-llm-jp-4\u3002&lt;think&gt; \u306e\u76ee\u5370\u304c\u7121\u3044\u306e\u3067\u3001\u305d\u306e\u307e\u307e\u3067\u306f\u601d\u8003\u304c\u56de\u7b54\u306b\u6df7\u3056\u308b)\u3002\n    &quot;&quot;&quot;\n    if &quot;&lt;|channel|&gt;&quot; in generated:\n        parts = _HARMONY_RE.findall(generated)\n        answer = &quot;&quot;.join(body for channel, body in parts if channel == &quot;final&quot;)\n        thinking = &quot;&quot;.join(body for channel, body in parts if channel != &quot;final&quot;)\n        return answer, thinking\n    text = generated\n    opened = None\n    tail = prompt_text.rstrip()\n    for m in _THINK_OPEN_RE.finditer(tail):\n        if m.end() == len(tail):\n            opened = m\n    if opened:\n        text = opened.group(0) + text\n    m = _THINK_OPEN_RE.search(text)\n    if not m:\n        return text, &quot;&quot;\n    before, rest = text[:m.start()], text[m.end():]\n    name = m.group(1)\n    closers = [f&quot;&lt;\/{name}&gt;&quot;]\n    if &quot;|&quot; in name or &quot;_&quot; in name:\n        # \u5909\u7a2e(think_fast)\u306e\u9589\u3058\u65b9\u304c\u672c\u4f53(&lt;\/ifm|think&gt;)\u3068\u540c\u3058\u30c6\u30f3\u30d7\u30ec\u30fc\u30c8\u304c\u3042\u308b\n        closers.append(&quot;&lt;\/&quot; + re.sub(r&quot;(think)[^|&lt;&gt;]*$&quot;, r&quot;\\1&quot;, name) + &quot;&gt;&quot;)\n    found = [(rest.find(c), c) for c in closers if c in rest]\n    if found:\n        at, closer = min(found)\n        return before + rest[at + len(closer):], rest[:at]\n    return before, rest\n\ndef run(job: dict) -&gt; dict:\n    &quot;&quot;&quot;\n    job: {&quot;dir&quot;, &quot;prompts&quot;: [{&quot;id&quot;, &quot;ja&quot;}], &quot;max_tokens&quot;}\u3002\n    \u623b\u308a\u5024\u306f\u8a18\u9332\u306e\u4e00\u90e8\u3002\u30c1\u30e3\u30c3\u30c8\u30c6\u30f3\u30d7\u30ec\u30fc\u30c8\u304c\u7121\u3051\u308c\u3070 {&quot;no_chat_template&quot;: True}\u3002\n    &quot;&quot;&quot;\n    import torch\n    import transformers\n    from transformers import AutoModelForCausalLM, AutoTokenizer\n    os.environ[&quot;HF_HUB_OFFLINE&quot;] = &quot;1&quot;\n    os.environ[&quot;TRANSFORMERS_OFFLINE&quot;] = &quot;1&quot;\n    t0 = time.time()\n    path = os.path.join(MOUNT, job[&quot;dir&quot;])\n    gpu = torch.cuda.get_device_name(0)\n    vram_total = torch.cuda.get_device_properties(0).total_memory\n    _log(t0, f&quot;GPU: {gpu}\u30fbtransformers {transformers.__version__}\u30fbtorch {torch.__version__}&quot;)\n    # \u7248\u306f str \u306b\u3059\u308b(torch.__version__ \u306f torch \u306e\u578b\u3067\u3001torch \u306e\u7121\u3044\u30db\u30b9\u30c8\u3067\u7d50\u679c\u3092\u8aad\u3081\u306a\u3044\u30022026-09-30\u30011\u672c\u3076\u3093\u5931\u3063\u305f)\n    record = {&quot;gpu&quot;: gpu, &quot;vram_total&quot;: int(vram_total), &quot;transformers&quot;: str(transformers.__version__),\n              &quot;torch&quot;: str(torch.__version__), &quot;dtype&quot;: &quot;bfloat16&quot;, &quot;packages&quot;: installed_packages(),\n              &quot;python&quot;: platform.python_version(),\n              &quot;measured_at&quot;: datetime.now(timezone.utc).isoformat()}\n    # \u91cf\u5b50\u5316\u6e08\u307f\u306e\u914d\u5e03\u306a\u3089\u3001\u305d\u306e\u5f62\u5f0f\u3092\u8a18\u9332\u3059\u308b(\u8a18\u4e8b\u306b\u300cBF16 \u306e\u307e\u307e\u300d\u3068\u66f8\u304b\u306a\u3044)\n    with open(os.path.join(path, &quot;config.json&quot;), encoding=&quot;utf-8&quot;) as f:\n        quant = json.load(f).get(&quot;quantization_config&quot;) or {}\n    if quant:\n        groups = list((quant.get(&quot;config_groups&quot;) or {}).values())\n        record[&quot;quantization&quot;] = {&quot;quant_method&quot;: quant.get(&quot;quant_method&quot;),\n                                  &quot;format&quot;: quant.get(&quot;format&quot;) or (groups[0].get(&quot;format&quot;) if groups else None),\n                                  &quot;weights_bits&quot;: ((groups[0].get(&quot;weights&quot;) or {}).get(&quot;num_bits&quot;) if groups else None)}\n        _log(t0, f&quot;\u91cf\u5b50\u5316\u6e08\u307f\u306e\u914d\u5e03: {record['quantization']}&quot;)\n    tok = AutoTokenizer.from_pretrained(path, trust_remote_code=True, local_files_only=True)\n    if not getattr(tok, &quot;chat_template&quot;, None):\n        _log(t0, &quot;\u30c1\u30e3\u30c3\u30c8\u30c6\u30f3\u30d7\u30ec\u30fc\u30c8\u304c\u3042\u308a\u307e\u305b\u3093&quot;)\n        return plain({**record, &quot;no_chat_template&quot;: True})\n    model = AutoModelForCausalLM.from_pretrained(path, trust_remote_code=True, local_files_only=True,\n                                                 use_safetensors=True, dtype=torch.bfloat16,\n                                                 device_map=&quot;cuda&quot;)\n    model.eval()\n    torch.cuda.synchronize()\n    # \u8aad\u307f\u8fbc\u3093\u3060\u5f8c\u306e\u91cd\u307f\u306e\u6301\u3061\u65b9(\u578b\u3054\u3068\u306e\u8981\u7d20\u6570\u3068\u3001\u91cd\u307f\u3092\u6301\u3064\u672b\u7aef\u306e\u5c64\u306e\u30af\u30e9\u30b9)\u3002\u91cf\u5b50\u5316\u6e08\u307f\u306e\u914d\u5e03\u304c 4bit \u306e\u307e\u307e\n    # \u8aad\u307e\u308c\u305f\u306e\u304b\u3001BF16 \u3078\u5c55\u958b\u3055\u308c\u305f\u306e\u304b\u3092\u3001\u63a8\u6e2c\u3067\u306f\u306a\u304f\u6570\u3048\u3066\u8a18\u9332\u3059\u308b\n    dtypes, classes = {}, {}\n    for tensor in list(model.parameters()) + list(model.buffers()):\n        name = str(tensor.dtype).replace(&quot;torch.&quot;, &quot;&quot;)\n        dtypes[name] = dtypes.get(name, 0) + int(tensor.numel())\n    for module in model.modules():\n        if not list(module.children()) and list(module.parameters(recurse=False)):\n            cls = type(module).__name__\n            classes[cls] = classes.get(cls, 0) + 1\n    record[&quot;loaded_weights&quot;] = {&quot;dtypes&quot;: dtypes, &quot;leaf_classes&quot;: classes}\n    record[&quot;vram_loaded&quot;] = int(torch.cuda.memory_allocated())\n    record[&quot;load_sec&quot;] = round(time.time() - t0)\n    _log(t0, f&quot;\u8aad\u307f\u8fbc\u307f: VRAM {record['vram_loaded'] \/ 1024 ** 3:.1f}GB\u30fb{record['load_sec']}\u79d2&quot;)\n    torch.cuda.reset_peak_memory_stats()\n    outputs = []\n    # \u6253\u3061\u5207\u308a(Modal \u306e timeout)\u306e\u524d\u306b\u3001\u7b54\u3048\u305f\u5206\u3060\u3051\u3067\u8fd4\u3059\u3002\u6253\u3061\u5207\u3089\u308c\u308b\u3068\u7b54\u3048\u305f\u5206\u3082\u5168\u3066\u5931\u3046\n    # \u6b21\u306e1\u554f\u304c\u6700\u9577(max_tokens)\u307e\u3067\u66f8\u3044\u305f\u3068\u304d\u306e\u6642\u9593\u3092\u3001\u3053\u3053\u307e\u3067\u3067\u6700\u3082\u9045\u304b\u3063\u305f\u901f\u3055\u3067\u898b\u7a4d\u3082\u3063\u3066\u5224\u5b9a\u3059\u308b\n    stop_at = t0 + job.get(&quot;answer_budget_sec&quot;, 10 ** 9)\n    for prompt in job[&quot;prompts&quot;]:\n        slowest = min((o[&quot;gen_tps&quot;] for o in outputs if o.get(&quot;gen_tps&quot;)), default=None)\n        worst = job[&quot;max_tokens&quot;] \/ slowest if slowest else 0\n        if time.time() + worst &gt; stop_at:\n            _log(t0, f&quot;\u6642\u9593\u306e\u4e0a\u9650\u304c\u8fd1\u3044\u305f\u3081\u3001\u6b8b\u308a\u306e\u554f\u3044({prompt['id']} \u304b\u3089)\u306f\u5c0b\u306d\u307e\u305b\u3093&quot;)\n            record[&quot;unanswered&quot;] = [p[&quot;id&quot;] for p in job[&quot;prompts&quot;][len(outputs):]]\n            break\n        ids = _chat_ids(tok, prompt[&quot;ja&quot;]).to(&quot;cuda&quot;)\n        started = time.time()\n        with torch.no_grad():\n            generated = model.generate(ids, max_new_tokens=job[&quot;max_tokens&quot;], do_sample=False)\n        torch.cuda.synchronize()\n        seconds = time.time() - started\n        new = generated[0][ids.shape[1]:]\n        count = int(new.shape[0])\n        # \u7279\u5225\u306a\u8a18\u53f7(&lt;think&gt; \u7b49)\u3092\u6b8b\u3057\u3066\u6587\u5b57\u306b\u3057\u3001\u601d\u8003\u3068\u56de\u7b54\u3092\u5206\u3051\u3066\u304b\u3089\u3001\u6b8b\u308a\u306e\u7279\u5225\u306a\u8a18\u53f7\u3092\u6d88\u3059\n        prompt_text = tok.decode(ids[0][-8:], skip_special_tokens=False)\n        answer, thinking = split_thinking(tok.decode(new, skip_special_tokens=False), prompt_text)\n        for special in control_tokens(tok):\n            answer = answer.replace(special, &quot;&quot;)\n        outputs.append({\n            &quot;id&quot;: prompt[&quot;id&quot;],\n            &quot;content&quot;: answer.strip(),\n            &quot;reasoning_chars&quot;: len(thinking.strip()),\n            &quot;finish_reason&quot;: &quot;length&quot; if count &gt;= job[&quot;max_tokens&quot;] else &quot;stop&quot;,\n            &quot;completion_tokens&quot;: count,\n            &quot;gen_tps&quot;: round(count \/ seconds, 2) if seconds &gt; 0 else None,\n            &quot;prompt_tokens&quot;: int(ids.shape[1]),\n        })\n        _log(t0, f&quot;\u56de\u7b54 {prompt['id']}: {count}\u30c8\u30fc\u30af\u30f3\u30fb{outputs[-1]['gen_tps']} \u30c8\u30fc\u30af\u30f3\/\u79d2&quot;)\n    record[&quot;vram_peak&quot;] = int(torch.cuda.max_memory_allocated())\n    record[&quot;outputs&quot;] = outputs\n    gen = sorted(o[&quot;gen_tps&quot;] for o in outputs if o.get(&quot;gen_tps&quot;))\n    if gen:\n        record[&quot;tg_tps&quot;] = gen[len(gen) \/\/ 2]\n    return plain(record)\n\n_CONTROL_RE = re.compile(r&quot;^&lt;\\|[^&lt;&gt;\\s]{1,40}?\\|&gt;$&quot;)\n\ndef control_tokens(tok) -&gt; list[str]:\n    &quot;&quot;&quot;\n    \u56de\u7b54\u304b\u3089\u5916\u3059\u30c6\u30f3\u30d7\u30ec\u30fc\u30c8\u306e\u8a18\u53f7\u3002\u7279\u5225\u306a\u8a18\u53f7(all_special_tokens)\u306b\u52a0\u3048\u3001\u8ffd\u52a0\u306e\u8a9e\u5f59\u306e\u3046\u3061 &lt;|\u2026|&gt; \u306e\u5f62\u306e\u3082\u306e\n    (K2-Horizon \u306e &lt;|ifm|im_end|&gt; \u306f\u7279\u5225\u306a\u8a18\u53f7\u306b\u767b\u9332\u3055\u308c\u3066\u304a\u3089\u305a\u3001\u56de\u7b54\u306e\u672b\u5c3e\u306b\u6b8b\u3063\u3066\u516c\u958b\u3055\u308c\u305f\u30022026-10-02)\u3002\n    \u9577\u3044\u3082\u306e\u304b\u3089\u5916\u3059(\u77ed\u3044\u8a18\u53f7\u304c\u9577\u3044\u8a18\u53f7\u306e\u4e00\u90e8\u3092\u5148\u306b\u6d88\u3055\u306a\u3044\u3088\u3046\u306b)\u3002\n    &quot;&quot;&quot;\n    found = set(tok.all_special_tokens)\n    found |= {t for t in (tok.get_added_vocab() or {}) if _CONTROL_RE.match(t)}\n    return sorted(found, key=len, reverse=True)\n\ndef plain(value):\n    &quot;&quot;&quot;JSON \u306b\u901a\u3057\u3066\u3001\u6a19\u6e96\u306e\u578b\u3060\u3051\u306b\u3059\u308b(\u30db\u30b9\u30c8\u306b\u306f torch \u3082 transformers \u3082\u7121\u3044)\u3002&quot;&quot;&quot;\n    return json.loads(json.dumps(value, default=str))\n<\/code><\/pre>\n<p>The <code>job<\/code> passed to <code>download<\/code> (as used):<\/p>\n<pre><code class=\"language-json\">{\n  &quot;repo&quot;: &quot;elyza\/ELYZA-Thinking-1.0-llm-jp-4-32b-a3b&quot;,\n  &quot;revision&quot;: &quot;5260ecc249f32ef5005ef940c78f41f130d9c468&quot;,\n  &quot;files&quot;: [\n    &quot;chat_template.jinja&quot;,\n    &quot;config.json&quot;,\n    &quot;generation_config.json&quot;,\n    &quot;llmjp4_harmony.py&quot;,\n    &quot;llmjp4_tokenizer.py&quot;,\n    &quot;model-00001-of-00002.safetensors&quot;,\n    &quot;model-00002-of-00002.safetensors&quot;,\n    &quot;model.safetensors.index.json&quot;,\n    &quot;tokenizer.json&quot;,\n    &quot;tokenizer_config.json&quot;\n  ],\n  &quot;dir&quot;: &quot;run-1791653766&quot;\n}\n<\/code><\/pre>\n<p>The <code>job<\/code> passed to <code>run<\/code> (as used):<\/p>\n<pre><code class=\"language-json\">{\n  &quot;dir&quot;: &quot;run-1791653766&quot;,\n  &quot;max_tokens&quot;: 1024,\n  &quot;answer_budget_sec&quot;: 3420,\n  &quot;prompts&quot;: [\n    {\n      &quot;id&quot;: &quot;summary&quot;,\n      &quot;ja&quot;: &quot;\u6b21\u306e\u8a18\u4e8b\u30923\u6587\u3067\u8981\u7d04\u3057\u3066\u304f\u3060\u3055\u3044\u3002\\n\\n\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\\n\\n\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\\n\\n\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\\n\\n\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\\n\\n\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\\n\\n\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\\n\\n\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\\n\\n\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&quot;\n    },\n    {\n      &quot;id&quot;: &quot;keigo&quot;,\n      &quot;ja&quot;: &quot;\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&quot;\n    },\n    {\n      &quot;id&quot;: &quot;arithmetic&quot;,\n      &quot;ja&quot;: &quot;\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&quot;\n    },\n    {\n      &quot;id&quot;: &quot;prefectures&quot;,\n      &quot;ja&quot;: &quot;\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&quot;\n    },\n    {\n      &quot;id&quot;: &quot;haiku&quot;,\n      &quot;ja&quot;: &quot;\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&quot;\n    }\n  ]\n}\n<\/code><\/pre>\n<details>\n<summary>Package versions in the running container (full list)<\/summary>\n<pre><code>accelerate==1.15.0 aiohappyeyeballs==2.6.1 aiohttp==3.12.7 aiosignal==1.3.2 annotated-doc==0.0.5 annotated-types==0.8.0 anyio==4.15.1 attrs==25.3.0 cbor2==5.7.0 certifi==2026.7.22 charset-normalizer==3.5.2 click==8.5.0 compressed-tensors==0.19.0 cuda-bindings==13.4.3 cuda-pathfinder==1.8.2 cuda-toolkit==13.0.3.0 einops==0.8.2 filelock==4.0.8 frozenlist==1.6.0 fsspec==2026.9.0 grpclib==0.4.8 h11==0.16.0 h2==4.2.0 hf-xet==1.6.0 hpack==4.1.0 httpcore==1.0.9 httpx==0.28.1 huggingface_hub==1.33.0 hyperframe==6.1.0 idna==3.20 Jinja2==3.1.6 loguru==0.7.3 markdown-it-py==4.2.0 MarkupSafe==3.0.3 mdurl==0.1.2 mpmath==1.3.0 multidict==6.4.4 networkx==3.7 numpy==2.5.3 nvidia-cublas==13.1.1.3 nvidia-cuda-cupti==13.0.85 nvidia-cuda-nvrtc==13.0.88 nvidia-cuda-runtime==13.0.96 nvidia-cudnn-cu13==9.24.0.43 nvidia-cufft==12.0.0.61 nvidia-cufile==1.15.1.6 nvidia-curand==10.4.0.35 nvidia-cusolver==12.0.4.66 nvidia-cusparse==12.6.3.3 nvidia-cusparselt-cu13==0.8.1 nvidia-nccl-cu13==2.30.7 nvidia-nvjitlink==13.4.92 nvidia-nvshmem-cu13==3.4.5 nvidia-nvtx==13.0.85 packaging==26.3 pip==25.1.1 propcache==0.3.1 protobuf==7.36.2 psutil==7.2.2 pydantic==2.13.5 pydantic_core==2.46.5 Pygments==2.21.0 PyYAML==6.0.3 regex==2026.9.29 requests==2.34.2 rich==15.0.0 safetensors==0.8.0 sentencepiece==0.2.2 setuptools==84.0.0 shellingham==1.5.4 sympy==1.14.0 tiktoken==0.14.0 tokenizers==0.23.2 torch==2.14.1 tqdm==4.70.1 transformers==5.18.0 triton==3.8.0 typer==0.27.2 typing-inspection==0.4.4 typing_extensions==4.16.0 urllib3==2.8.0 uv==0.7.19 wheel==0.45.1 yarl==1.20.0<\/code><\/pre>\n<\/details>\n<details>\n<summary>Package versions in the downloading container (full list)<\/summary>\n<pre><code>aiohappyeyeballs==2.6.1 aiohttp==3.12.7 aiosignal==1.3.2 anyio==4.15.1 attrs==25.3.0 cbor2==5.7.0 certifi==2025.4.26 click==8.5.0 filelock==4.0.7 frozenlist==1.6.0 fsspec==2026.9.0 grpclib==0.4.8 h11==0.16.0 h2==4.2.0 hf-xet==1.6.0 hpack==4.1.0 httpcore2==2.13.1 httpx2==2.13.1 huggingface_hub==2.0.0 hyperframe==6.1.0 idna==3.20 multidict==6.4.4 packaging==26.3 pip==25.1.1 propcache==0.3.1 protobuf==6.31.1 PyYAML==6.0.3 tqdm==4.70.1 truststore==0.10.4 typing_extensions==4.16.0 uv==0.7.19 wheel==0.45.1 yarl==1.20.0<\/code><\/pre>\n<\/details>\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>LiquidAI\/LFM2.5-24B-A2B-Base<\/td>\n<td>24.0B<\/td>\n<td>80GB<\/td>\n<td>lfm1.0<\/td>\n<td><a href=\"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/2026\/11\/17\/liquidai-lfm2-5-24b-a2b-2\/\">LiquidAI Releases LFM2.5-24B-A2B for On-Device Agents<\/a> (2026-11-17)<\/td>\n<\/tr>\n<tr>\n<td>LiquidAI\/LFM2.5-24B-A2B-FP8<\/td>\n<td>24.0B<\/td>\n<td>32GB<\/td>\n<td>lfm1.0<\/td>\n<td><a href=\"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/2026\/11\/17\/liquidai-lfm2-5-24b-a2b-fp8-released\/\">LFM2.5-24B-A2B-FP8 Text Generation Model: 32GB+ VRAM<\/a> (2026-11-17)<\/td>\n<\/tr>\n<tr>\n<td>LiquidAI\/LFM2.5-24B-A2B-MLX-bf16<\/td>\n<td>24.0B<\/td>\n<td>80GB<\/td>\n<td>lfm1.0<\/td>\n<td><a href=\"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/2026\/11\/17\/liquidai-lfm2-5-24b-a2b-mlx-bf16-2\/\">LFM2.5-24B-A2B-MLX-bf16 Reasoning Model for Apple Silicon: 80GB+ VRAM<\/a> (2026-11-17)<\/td>\n<\/tr>\n<tr>\n<td>LiquidAI\/LFM2.5-24B-A2B-MLX-8bit<\/td>\n<td>24.0B<\/td>\n<td>32GB<\/td>\n<td>lfm1.0<\/td>\n<td><a href=\"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/2026\/11\/17\/liquidai-lfm2-5-24b-a2b-mlx-8bit\/\">LFM2.5-24B-A2B-MLX-8bit Text Generation Model: 32GB+ VRAM<\/a> (2026-11-17)<\/td>\n<\/tr>\n<tr>\n<td>LiquidAI\/LFM2.5-24B-A2B<\/td>\n<td>24.0B<\/td>\n<td>12GB<\/td>\n<td>lfm1.0<\/td>\n<td><a href=\"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/2026\/11\/17\/liquidai-lfm25-24b-a2b\/\">LFM2.5-24B-A2B Agent-Optimized MoE Reasoning Model: 12GB+ VRAM<\/a> (2026-11-17)<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<\/div>\n<p><!-- \/lmw:peers --><\/p>\n<h2>How to Get It<\/h2>\n<p>The model is distributed in safetensors format and is available on Hugging Face under the Apache-2.0 license. There are no access restrictions or gated license agreements required on Hugging Face to use it.<\/p>\n<p>It is compatible with engines such as vLLM and Transformers, and an example command for launching a server using vLLM is as follows:<\/p>\n<pre><code class=\"language-bash\">vllm serve elyza\/ELYZA-Thinking-1.0-llm-jp-4-32b-a3b \\\n  --max-model-len 65536 \\\n  --trust-remote-code \\\n  --moe-backend triton \\\n  --reasoning-parser-plugin vllm_plugins\/multi_parser_loader.py \\\n  --reasoning-parser llmjp4 \\\n  --tool-parser-plugin vllm_plugins\/llmjp_harmony_tool_parser.py \\\n  --tool-call-parser llmjp_harmony \\\n  --enable-auto-tool-choice\n<\/code><\/pre>\n<p><!-- lmw:same-task --><\/p>\n<h2>Other Models for the Same Task<\/h2>\n<p><em>Recent text generation models covered by Local Model Watch, newest first. Grouped by the task each publisher declares on Hugging Face (pipeline_tag); the smallest VRAM tier is this site&#8217;s estimate.<\/em><\/p>\n<ul>\n<li><a href=\"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/2026\/11\/17\/plamo-3-8b-fin-instruct-2\/\">plamo-3-8b-fin-instruct Text Generation Model: 4GB+ VRAM, GGUF Builds<\/a> (8.5B, 4GB)<\/li>\n<li><a href=\"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/2026\/11\/17\/tencent-qwen3-8-27b-dspine-b16\/\">Qwen3.8-27B-DSpine-b16 Draft Model for Speculative Decoding: 8GB+ VRAM<\/a> (2.0B, 8GB)<\/li>\n<li><a href=\"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/2026\/11\/17\/arete-27b-released\/\">Arete-27B Text Generation Model: 12GB+ VRAM, GGUF Builds<\/a> (27.4B, 12GB)<\/li>\n<li><a href=\"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/2026\/11\/17\/bytedance-seed-releases-cola2-language-model\/\">ByteDance-Seed Releases Cola2 Language Model<\/a> (8.8B, 48GB)<\/li>\n<\/ul>\n<p><a href=\"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/models-by-task-en\/#task-text\">See all text generation models \u2192<\/a><\/p>\n<p><!-- \/lmw:same-task --><\/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 needs at least 80GB) \u2192 <a href=\"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/vram-guide-en\/\">VRAM quick reference<\/a><\/li>\n<li><strong>Engines that run this model<\/strong> \u2192 <a href=\"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/engine-vllm-en\/\">vLLM<\/a><\/li>\n<li><strong>Formats this model is available in<\/strong> \u2192 <a href=\"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/format-safetensors-en\/\">Safetensors format guide and models<\/a><\/li>\n<li><strong>Learn about the publisher<\/strong> \u2192 <a href=\"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/publisher-elyza-en\/\">ELYZA: models, licenses and articles<\/a><\/li>\n<li><strong>How to read MMLU-Pro, GPQA Diamond, MATH<\/strong> \u2192 <a href=\"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/glossary-benchmarks-en\/\">Benchmark glossary<\/a><\/li>\n<\/ul>\n<p><!-- \/lmw:next-steps --><\/p>\n<h2>Sources<\/h2>\n<ul>\n<li><a href=\"https:\/\/huggingface.co\/elyza\/ELYZA-Thinking-1.0-llm-jp-4-32b-a3b\">https:\/\/huggingface.co\/elyza\/ELYZA-Thinking-1.0-llm-jp-4-32b-a3b<\/a><\/li>\n<li><a href=\"https:\/\/huggingface.co\/elyza\/ELYZA-Thinking-1.0-llm-jp-4-33b\">https:\/\/huggingface.co\/elyza\/ELYZA-Thinking-1.0-llm-jp-4-33b<\/a><\/li>\n<li><a href=\"https:\/\/huggingface.co\/llm-jp\/llm-jp-4-33b-base\">https:\/\/huggingface.co\/llm-jp\/llm-jp-4-33b-base<\/a><\/li>\n<\/ul>\n<p><!-- lmw:updates --><\/p>\n<h2>Update History<\/h2>\n<ul>\n<li>2026-10-03: Added our own measurements: Japanese token efficiency.<\/li>\n<li>2026-10-11: Added our own measurements: VRAM and answers to Japanese questions with the publisher&#8217;s own code.<\/li>\n<\/ul>\n<p><!-- \/lmw:updates --><\/p>\n","protected":false},"excerpt":{"rendered":"<p>ELYZA has released ELYZA-Thinking-1.0, a reasoning model based on the llm-jp-4 series, featuring strong Japanese-English bilingual capabilities.<\/p>\n","protected":false},"author":1,"featured_media":9040,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[310],"tags":[2884,2886,2888,165,1547,1555],"class_list":["post-9041","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-new-models","tag-elyza-en","tag-elyza-thinking-1-0-en","tag-llm-jp-4-en","tag-moe-en","tag-verified","tag--en"],"lang":"en","translations":{"en":9041,"ja":9039},"_links":{"self":[{"href":"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/wp-json\/wp\/v2\/posts\/9041","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=9041"}],"version-history":[{"count":54,"href":"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/wp-json\/wp\/v2\/posts\/9041\/revisions"}],"predecessor-version":[{"id":17665,"href":"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/wp-json\/wp\/v2\/posts\/9041\/revisions\/17665"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/wp-json\/wp\/v2\/media\/9040"}],"wp:attachment":[{"href":"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/wp-json\/wp\/v2\/media?parent=9041"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/wp-json\/wp\/v2\/categories?post=9041"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/wp-json\/wp\/v2\/tags?post=9041"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}