{"id":627,"date":"2026-09-14T20:34:33","date_gmt":"2026-09-14T11:34:33","guid":{"rendered":"https:\/\/localmodelwatch.tsuchitsuchi.com\/2026\/09\/14\/orion-26b-a4b-v1-1-gguf-quantizations\/"},"modified":"2026-10-01T08:12:51","modified_gmt":"2026-09-30T23:12:51","slug":"orion-26b-a4b-v1-1-gguf-quantizations","status":"publish","type":"post","link":"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/2026\/09\/14\/orion-26b-a4b-v1-1-gguf-quantizations\/","title":{"rendered":"TheDrummer_Orion-26B-A4B-v1.1-GGUF Multimodal Model: 12GB+ 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\/bartowski\/TheDrummer_Orion-26B-A4B-v1.1-GGUF\">bartowski\/TheDrummer_Orion-26B-A4B-v1.1-GGUF<\/a><\/td>\n<\/tr>\n<tr>\n<td>Published<\/td>\n<td>2026-09-14<\/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>It needs about 18% fewer tokens than the Qwen3 tokenizer for the same Japanese text, so about 1.22\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>It is reported that bartowski has released GGUF quantization files for the multimodal model Orion-26B-A4B-v1.1. As some information included could not be officially verified at the time of writing, caution is advised when deploying or testing the model in practice. It supports text, image, and audio inputs, and a variety of quantization files are provided for local execution.<\/p>\n<h2>Specifications<\/h2>\n<ul>\n<li>Parameters: 26B<\/li>\n<li>Input Support: Text, image, audio (requires mmproj file)<\/li>\n<li>Speculative Decoding: None<\/li>\n<li>imatrix: Supported<\/li>\n<li>Architecture: Gemma4ForConditionalGeneration<\/li>\n<\/ul>\n<h2>Performance<\/h2>\n<p>Regarding the performance of the base model, according to the model card, a low temperature setting (such as 0.9) is optimal, and Gemma&#8217;s thinking capabilities are reported to be very sharp in roleplay (RP). Below is a comparison table of per-tensor layouts for the published quantized versions.<\/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>Quant<\/th>\n<th>Size<\/th>\n<th>Body bits\/weight<\/th>\n<th>File bits\/weight<\/th>\n<th>Body kept at base type<\/th>\n<\/tr>\n<\/thead>\n<tbody>\n<tr>\n<td>Q6_K_L<\/td>\n<td>24.52GB<\/td>\n<td>7.38<\/td>\n<td>7.60<\/td>\n<td>51 %<\/td>\n<\/tr>\n<tr>\n<td>Q6_K<\/td>\n<td>23.79GB<\/td>\n<td>7.03<\/td>\n<td>7.37<\/td>\n<td>71 %<\/td>\n<\/tr>\n<tr>\n<td>Q6_K_S<\/td>\n<td>23.14GB<\/td>\n<td>6.71<\/td>\n<td>7.17<\/td>\n<td>90 %<\/td>\n<\/tr>\n<tr>\n<td>Q5_K_M<\/td>\n<td>20.01GB<\/td>\n<td>5.92<\/td>\n<td>6.20<\/td>\n<td>71 %<\/td>\n<\/tr>\n<tr>\n<td>Q5_K_S<\/td>\n<td>18.40GB<\/td>\n<td>5.70<\/td>\n<td>5.70<\/td>\n<td>90 %<\/td>\n<\/tr>\n<tr>\n<td>Q4_K_L<\/td>\n<td>18.21GB<\/td>\n<td>5.19<\/td>\n<td>5.64<\/td>\n<td>50 %<\/td>\n<\/tr>\n<tr>\n<td>Q4_K_M<\/td>\n<td>17.97GB<\/td>\n<td>5.07<\/td>\n<td>5.57<\/td>\n<td>70 %<\/td>\n<\/tr>\n<tr>\n<td>Q4_K_S<\/td>\n<td>16.05GB<\/td>\n<td>4.70<\/td>\n<td>4.98<\/td>\n<td>90 %<\/td>\n<\/tr>\n<tr>\n<td>IQ4_NL<\/td>\n<td>15.63GB<\/td>\n<td>5.07<\/td>\n<td>4.85<\/td>\n<td>70 %<\/td>\n<\/tr>\n<tr>\n<td>Q3_K_L<\/td>\n<td>14.80GB<\/td>\n<td>4.39<\/td>\n<td>4.59<\/td>\n<td>50 %<\/td>\n<\/tr>\n<tr>\n<td>IQ3_M<\/td>\n<td>14.34GB<\/td>\n<td>4.17<\/td>\n<td>4.45<\/td>\n<td>50 %<\/td>\n<\/tr>\n<tr>\n<td>IQ4_XS<\/td>\n<td>14.28GB<\/td>\n<td>4.42<\/td>\n<td>4.43<\/td>\n<td>90 %<\/td>\n<\/tr>\n<tr>\n<td>Q3_K_M<\/td>\n<td>13.79GB<\/td>\n<td>4.01<\/td>\n<td>4.27<\/td>\n<td>70 %<\/td>\n<\/tr>\n<tr>\n<td>IQ3_XS<\/td>\n<td>12.84GB<\/td>\n<td>3.57<\/td>\n<td>3.98<\/td>\n<td>90 %<\/td>\n<\/tr>\n<tr>\n<td>IQ3_XXS<\/td>\n<td>12.61GB<\/td>\n<td>3.46<\/td>\n<td>3.91<\/td>\n<td>70 %<\/td>\n<\/tr>\n<tr>\n<td>Q3_K_S<\/td>\n<td>12.37GB<\/td>\n<td>3.56<\/td>\n<td>3.84<\/td>\n<td>90 %<\/td>\n<\/tr>\n<tr>\n<td>Q2_K<\/td>\n<td>11.61GB<\/td>\n<td>3.19<\/td>\n<td>3.60<\/td>\n<td>70 %<\/td>\n<\/tr>\n<tr>\n<td>IQ2_M<\/td>\n<td>10.90GB<\/td>\n<td>2.89<\/td>\n<td>3.38<\/td>\n<td>70 %<\/td>\n<\/tr>\n<tr>\n<td>IQ2_S<\/td>\n<td>10.31GB<\/td>\n<td>2.61<\/td>\n<td>3.20<\/td>\n<td>70 %<\/td>\n<\/tr>\n<tr>\n<td>IQ2_XS<\/td>\n<td>9.90GB<\/td>\n<td>2.41<\/td>\n<td>3.07<\/td>\n<td>90 %<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<\/div>\n<p>(Table truncated. Subsequent rows omitted)<\/p>\n<p>Furthermore, the comparative verification results between the computed layout and the standard layout are shown below.<\/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>Quant<\/th>\n<th>Computed layout KLD<\/th>\n<th>Standard layout KLD<\/th>\n<th>Ratio at equal size<\/th>\n<th>Size vs standard file<\/th>\n<\/tr>\n<\/thead>\n<tbody>\n<tr>\n<td>Q4_K_M<\/td>\n<td>0.8086 \u00b1 0.0113<\/td>\n<td>1.1462 \u00b1 0.0132<\/td>\n<td>0.88\u00d7<\/td>\n<td>+5.5 %<\/td>\n<\/tr>\n<tr>\n<td>Q3_K_M<\/td>\n<td>1.2642 \u00b1 0.0149<\/td>\n<td>2.9859 \u00b1 0.0241<\/td>\n<td>0.51\u00d7<\/td>\n<td>+5.9 %<\/td>\n<\/tr>\n<tr>\n<td>IQ2_XS<\/td>\n<td>2.6730 \u00b1 0.0206<\/td>\n<td>2.9355 \u00b1 0.0210<\/td>\n<td>0.91\u00d7<\/td>\n<td>\u22122.3 %<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<\/div>\n<p>From these tables, it is confirmed that quantization files adopting the computed layout, such as Q4_K_M and Q3_K_M, keep the Kullback-Leibler divergence (KLD) lower compared to the standard layout, resulting in smaller differences from the unquantized model and maintained quality. Q4_K_M in particular shows a favorable balance.<\/p>\n<h2>Strengths and Use Cases<\/h2>\n<p>This model supports image and audio input in addition to text, and is intended for conversational use and roleplay (RP) applications. Inheriting the characteristics of the base model, it is well-suited for interactive tasks leveraging multimodal inputs.<\/p>\n<p><!-- lmw:hardware --><\/p>\n<h2>Hardware Requirements<\/h2>\n<p><strong>Estimated requirements (calculated by Local Model Watch)<\/strong> \u2014 25.8B parameters (taken from the base model TheDrummer\/Orion-26B-A4B-v1.1)<\/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>12GB (RTX 4070 \/ 3060 12GB, etc.)<\/td>\n<td>IQ2_S<\/td>\n<td>9.6GB<\/td>\n<td>11.5GB<\/td>\n<\/tr>\n<tr>\n<td>16GB (RTX 5060 Ti 16GB \/ 4060 Ti 16GB, etc.)<\/td>\n<td>IQ3_M<\/td>\n<td>13.4GB<\/td>\n<td>16.0GB<\/td>\n<\/tr>\n<tr>\n<td>24GB (RTX 4090 \/ 3090, etc.)<\/td>\n<td>Q5_K_M<\/td>\n<td>18.6GB<\/td>\n<td>22.4GB<\/td>\n<\/tr>\n<tr>\n<td>32GB (RTX 5090, etc.)<\/td>\n<td>Q8_0<\/td>\n<td>25.0GB<\/td>\n<td>30.0GB<\/td>\n<\/tr>\n<tr>\n<td>80GB class (A100 \/ H100)<\/td>\n<td>BF16<\/td>\n<td>47.0GB<\/td>\n<td>56.4GB<\/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>Compression:<\/strong> the IQ2_S build measures 3.2 bits per weight \u2014 about 20% the size of the original 16-bit weights, calculated by this site from the actual file sizes.<\/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.<\/em><\/p>\n<p><!-- \/lmw:runnability --><\/p>\n<p><!-- lmw:lab --><\/p>\n<h2>Our Own Measurements<\/h2>\n<p>Values we checked ourselves on our server (no GPU) by actually reading this model&#8217;s files, without running the model \u2014 not figures copied from the model card. How we measure, and the results for every model: <a href=\"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/observations-en\/\">Our Measurements<\/a>.<\/p>\n<h3>Measurements We Did Not Take<\/h3>\n<p>This model does not state a license, so we cannot confirm permission to use it. Because this site carries advertising, we did not run the model (no CPU run, answers, quantization comparison, conversion or generation). Only values we checked without running the model, such as token counts and the GGUF header, are shown.<\/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>564<\/strong><\/td>\n<td><strong>1.03\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 18% fewer tokens than the Qwen3 tokenizer for the same Japanese text, so about 1.22\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\/TheDrummer\/Orion-26B-A4B-v1.1\">TheDrummer\/Orion-26B-A4B-v1.1<\/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\/bartowski\/TheDrummer_Orion-26B-A4B-v1.1-GGUF\/blob\/main\/TheDrummer_Orion-26B-A4B-v1.1-Q4_K_M.gguf\">TheDrummer_Orion-26B-A4B-v1.1-Q4_K_M.gguf<\/a> (16.73GB, <code>Q4_K_M<\/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>gemma4<\/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>30<\/td>\n<\/tr>\n<tr>\n<td>Experts<\/td>\n<td>8 active out of 128<\/td>\n<\/tr>\n<tr>\n<td>Vocabulary size<\/td>\n<td>262,144<\/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>Used (<code>TheDrummer_Orion-26B-A4B-v1.1-calibration-v6.txt<\/code>, 492 chunks)<\/td>\n<\/tr>\n<tr>\n<td>Weight types (share of parameters)<\/td>\n<td>Q4_K 52.3% \/ Q8_0 20.2% \/ Q5_0 16.5% \/ Q6_K 9.0% \/ other 2.1%<\/td>\n<\/tr>\n<tr>\n<td>Average bits per weight<\/td>\n<td>5.69 bits<\/td>\n<\/tr>\n<tr>\n<td>Embedding \/ output layer type<\/td>\n<td>Q6_K \/ shared with the embedding<\/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<p><!-- \/lmw:lab --><\/p>\n<h2>How to Get It<\/h2>\n<ul>\n<li>Distribution Format: GGUF<\/li>\n<li>Supported Engines: llama.cpp, LM Studio, koboldcpp, ramalama, Jan AI, Text Generation Web UI, LoLLMs, Atomic Chat<\/li>\n<li>Example Download Command:<\/li>\n<\/ul>\n<pre><code>hf download bartowski\/TheDrummer_Orion-26B-A4B-v1.1-GGUF --include &quot;TheDrummer_Orion-26B-A4B-v1.1-Q4_K_M.gguf&quot; --local-dir.\/\n<\/code><\/pre>\n<p>When utilizing multimodal inputs, you will also need to prepare the corresponding multimodal projector file (<code>mmproj-TheDrummer_Orion-26B-A4B-v1.1-f16.gguf<\/code> or <code>mmproj-TheDrummer_Orion-26B-A4B-v1.1-bf16.gguf<\/code>).<\/p>\n<p><!-- lmw:related --><\/p>\n<h2>Related Articles<\/h2>\n<ul>\n<li><a href=\"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/2026\/09\/15\/salience-27b-r6-gguf-2\/\">vectionlabs_Salience-27B-R6-GGUF: Our Test Answers, 12GB+ VRAM<\/a><\/li>\n<li><a href=\"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/2026\/09\/15\/intern-s2-397b-gguf-released\/\">Intern-S2-397B-GGUF Vision-Language Model: ~102GB Memory<\/a><\/li>\n<li><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><\/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 12GB tier) \u2192 <a href=\"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/vram-guide-12gb-en\/\">Other models that run on a 12GB GPU<\/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 IQ2_S, IQ3_M, Q5_K_M 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\/glossary-quantization-en\/\">Quantization and model-format glossary<\/a><\/li>\n<li><strong>Other models for the same task<\/strong> \u2192 <a href=\"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/models-by-task-en\/#task-vision\">Other vision-language models<\/a><\/li>\n<\/ul>\n<p><!-- \/lmw:next-steps --><\/p>\n<h2>Sources<\/h2>\n<ul>\n<li><a href=\"https:\/\/huggingface.co\/bartowski\/TheDrummer_Orion-26B-A4B-v1.1-GGUF\">https:\/\/huggingface.co\/bartowski\/TheDrummer_Orion-26B-A4B-v1.1-GGUF<\/a><\/li>\n<li><a href=\"https:\/\/huggingface.co\/TheDrummer\/Orion-26B-A4B-v1.1\">https:\/\/huggingface.co\/TheDrummer\/Orion-26B-A4B-v1.1<\/a><\/li>\n<li><a href=\"https:\/\/github.com\/ggml-org\/llama.cpp\">https:\/\/github.com\/ggml-org\/llama.cpp<\/a><\/li>\n<li><a href=\"https:\/\/github.com\/bartowski1182\/quantization-config\">https:\/\/github.com\/bartowski1182\/quantization-config<\/a><\/li>\n<li><a href=\"https:\/\/huggingface.co\/blog\/bartowski\/per-tensor-layout-maps-for-gguf-quantization\">https:\/\/huggingface.co\/blog\/bartowski\/per-tensor-layout-maps-for-gguf-quantization<\/a><\/li>\n<li><a href=\"https:\/\/gist.github.com\/bartowski1182\/e26453c0404e24eb317543ec5360f87a\">https:\/\/gist.github.com\/bartowski1182\/e26453c0404e24eb317543ec5360f87a<\/a><\/li>\n<li><a href=\"https:\/\/github.com\/ggml-org\/llama.cpp\/pull\/9921\">https:\/\/github.com\/ggml-org\/llama.cpp\/pull\/9921<\/a><\/li>\n<li><a href=\"https:\/\/gist.github.com\/Artefact2\/b5f810600771265fc1e39442288e8ec9\">https:\/\/gist.github.com\/Artefact2\/b5f810600771265fc1e39442288e8ec9<\/a><\/li>\n<\/ul>\n<p><!-- lmw:updates --><\/p>\n<h2>Update History<\/h2>\n<ul>\n<li>2026-09-15: Verified the content against the official primary source.<\/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, what is inside the GGUF file.<\/li>\n<\/ul>\n<p><!-- \/lmw:updates --><\/p>\n","protected":false},"excerpt":{"rendered":"<p>Explore bartowski&#8217;s GGUF quantizations for Orion-26B-A4B-v1.1, featuring multimodal support, performance tables, and hardware requirements.<\/p>\n","protected":false},"author":1,"featured_media":3998,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[310],"tags":[1030,163,1073,1178,896,117],"class_list":["post-627","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-new-models","tag-bartowski-en","tag-gguf-en","tag-llama-cpp-en","tag-orion-26b-a4b-v1-1-en","tag--en"],"lang":"en","translations":{"en":627,"ja":625},"_links":{"self":[{"href":"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/wp-json\/wp\/v2\/posts\/627","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=627"}],"version-history":[{"count":28,"href":"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/wp-json\/wp\/v2\/posts\/627\/revisions"}],"predecessor-version":[{"id":8495,"href":"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/wp-json\/wp\/v2\/posts\/627\/revisions\/8495"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/wp-json\/wp\/v2\/media\/3998"}],"wp:attachment":[{"href":"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/wp-json\/wp\/v2\/media?parent=627"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/wp-json\/wp\/v2\/categories?post=627"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/wp-json\/wp\/v2\/tags?post=627"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}