{"id":13753,"date":"2026-11-17T11:02:31","date_gmt":"2026-11-17T02:02:31","guid":{"rendered":"https:\/\/localmodelwatch.tsuchitsuchi.com\/?p=13753"},"modified":"2026-11-17T11:11:42","modified_gmt":"2026-11-17T02:11:42","slug":"ling-3-1-flash-1019-released","status":"publish","type":"post","link":"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/2026\/11\/17\/ling-3-1-flash-1019-released\/","title":{"rendered":"Ling-3.1-flash-1019 Text Generation Model: ~641GB Memory, GGUF Builds"},"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\/inclusionAI\/Ling-3.1-flash-1019\">inclusionAI\/Ling-3.1-flash-1019<\/a><\/td>\n<\/tr>\n<tr>\n<td>Family guide<\/td>\n<td><a href=\"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/model-inclusionai-ling-3-1-en\/\">Ling-3.1 guide (3 articles)<\/a><\/td>\n<\/tr>\n<tr>\n<td>Publisher guide<\/td>\n<td><a href=\"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/publisher-inclusionai-en\/\">inclusionAI (Ant Group): models and licenses<\/a><\/td>\n<\/tr>\n<tr>\n<td>Published<\/td>\n<td>2026-10-27<\/td>\n<\/tr>\n<tr>\n<td>License<\/td>\n<td>mit<\/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<h2>Overview<\/h2>\n<p>inclusionAI has released the language model &#8220;Ling-3.1-flash-1019&#8221;. This model is an upgraded version of the existing Ling-3.1-flash model, featuring a hybrid linear MoE architecture with a total parameter count of 560B and 25B active parameters per token, supporting an ultra-long context window of 1M tokens.<\/p>\n<p>According to the model card, this model has undergone further post-training optimization, significantly improving its reasoning capabilities and real-world agent functions. It is designed to deliver high performance in multi-step task execution and agent workflows.<\/p>\n<h2>Specifications<\/h2>\n<ul>\n<li><strong>Architecture<\/strong>: Hybrid-linear MoE (BailingMoeV3ForCausalLM)<\/li>\n<li><strong>Total Parameters<\/strong>: 560B<\/li>\n<li><strong>Active Parameters<\/strong>: 25B<\/li>\n<li><strong>Context Length<\/strong>: 1M tokens<\/li>\n<li><strong>Transformer Layer Structure<\/strong>: 63 KDA + 9 Gated MLA (7:1)<\/li>\n<li><strong>Dense Layers<\/strong>: 4<\/li>\n<li><strong>Routing Experts<\/strong>: 512<\/li>\n<li><strong>Shared Experts<\/strong>: 1<\/li>\n<li><strong>Active Experts<\/strong>: 8<\/li>\n<li><strong>Attention Heads<\/strong>: 64<\/li>\n<li><strong>Hidden Size<\/strong>: 4096<\/li>\n<li><strong>Expert Intermediate Size<\/strong>: 1280<\/li>\n<li><strong>Dense Intermediate Size<\/strong>: 10240<\/li>\n<li><strong>Vocab Size<\/strong>: 157184<\/li>\n<\/ul>\n<h2>Performance<\/h2>\n<p>According to the measurement results shown in the model card, Ling-3.1-flash-1019 shows notable score improvements in both reasoning power and agent task evaluations compared to its predecessor, Ling-3.1-flash.<\/p>\n<p>The comparison results with the predecessor model published by the release source are 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>Evaluation Metric \/ Benchmark<\/th>\n<th>Predecessor Model (Ling-3.1-flash)<\/th>\n<th>Ling-3.1-flash-1019<\/th>\n<\/tr>\n<\/thead>\n<tbody>\n<tr>\n<td>Artificial Analysis Intelligence Index v4.3.2<\/td>\n<td>41<\/td>\n<td>44<\/td>\n<\/tr>\n<tr>\n<td>AA-Briefcase (Elo)<\/td>\n<td>Approx. 1400<\/td>\n<td>1524<\/td>\n<\/tr>\n<tr>\n<td>AutomationBench-AA<\/td>\n<td>62%<\/td>\n<td>68%<\/td>\n<\/tr>\n<tr>\n<td>CritPt<\/td>\n<td>18%<\/td>\n<td>26%<\/td>\n<\/tr>\n<tr>\n<td>SciCode<\/td>\n<td>54%<\/td>\n<td>57%<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<\/div>\n<p>From the description in the model card and the measurement scores above, the following trends can be observed.<\/p>\n<p>First, the overall intelligence index, &#8220;Artificial Analysis Intelligence Index v4.3.2,&#8221; improved its score from 41 to 44. In addition, the Elo score for AA-Briefcase increased from approximately 1400 to 1524, and AutomationBench-AA grew from 62% to 68%, indicating significant performance gains in autonomous workflows and complex task automation processing.<\/p>\n<p>Furthermore, improvements are seen in science and technology benchmarks such as CritPt (expanding from 18% to 26%) and SciCode (expanding from 54% to 57%). This indicates that the model possesses superior execution capabilities compared to its predecessor even in tasks involving advanced reasoning and specialized calculations or code processing.<\/p>\n<p>Note that these scores are not independently verified by a third-party organization, but are figures measured and published on the model card by the publisher itself, inclusionAI.<\/p>\n<h2>Strengths and Use Cases<\/h2>\n<p>Ling-3.1-flash-1019 is a model that enhances reasoning performance and real-world agent execution capabilities through post-training optimization. It is mainly suited for the following use cases and scenarios:<\/p>\n<ul>\n<li><strong>Agent Workflows and Multi-Step Task Execution<\/strong>: Excels in autonomous processing of tasks involving complex procedures and multi-step execution processes. The performance improvements seen in evaluations such as AutomationBench-AA and AA-Briefcase demonstrate high practicality as an autonomous agent.<\/li>\n<li><strong>Ultra-Long Document Processing and Context Utilization<\/strong>: Featuring a vast context window extending up to 1M tokens, it is suited for analyzing lengthy document groups, parsing large-scale repositories, and maintaining dialogues over extended session histories.<\/li>\n<li><strong>Tool Calling and Structured Reasoning<\/strong>: Supports external tool calling functions (tool calling) and the ability to unfold explicit thought processes within the model (thinking function). It can be integrated into advanced problem-solving workflows through parser configuration when launching the inference server.<\/li>\n<li><strong>Scientific Computation and Technical Code Generation<\/strong>: Based on improved benchmark scores such as CritPt and SciCode, usefulness can be expected in problem-solving and specialized coding tasks within the science and technology fields.<\/li>\n<\/ul>\n<p><!-- lmw:hardware --><\/p>\n<h2>Hardware Requirements<\/h2>\n<p><strong>Estimated requirements (calculated by Local Model Watch)<\/strong> \u2014 572.1B 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>More than 641GB of VRAM (multi-GPU or CPU offload required)<\/td>\n<td>FP8<\/td>\n<td>534.4GB<\/td>\n<td>641.3GB<\/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>Runs in Ollama, LM Studio and llama.cpp via a converted build.<\/strong><\/p>\n<p>The publisher ships safetensors, but <a href=\"https:\/\/huggingface.co\/bartowski\/Ling-3.1-flash-1019-GGUF\">bartowski\/Ling-3.1-flash-1019-GGUF<\/a> provides a GGUF build you can use.<\/p>\n<p><strong>License \u2014 <code>mit<\/code> (Commercial use allowed):<\/strong> Permits commercial use, modification and redistribution, provided the copyright notice and license text are retained.<\/p>\n<p><strong>Compression:<\/strong> the FP8 build measures 8.02 bits per weight \u2014 about 50% 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. 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:peers --><\/p>\n<h2>Recent Models in the Same Size Class<\/h2>\n<p><em>Models with <\/em><em>over 40B<\/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>inclusionAI\/Ling-3.1-flash<\/td>\n<td>572.1B<\/td>\n<td>\u2014<\/td>\n<td>mit<\/td>\n<td><a href=\"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/2026\/11\/17\/ling-31-flash-released\/\">Ling-3.1-flash Hybrid Reasoning MoE Model: ~641GB Memory, GGUF Builds<\/a> (2026-11-17)<\/td>\n<\/tr>\n<tr>\n<td>inclusionAI\/Ling-3.1-flash-base<\/td>\n<td>572.1B<\/td>\n<td>\u2014<\/td>\n<td>mit<\/td>\n<td><a href=\"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/2026\/11\/17\/ling-3-1-flash-base-released\/\">Ling-3.1-flash-base Released: 560B Hybrid Linear MoE Model<\/a> (2026-11-17)<\/td>\n<\/tr>\n<tr>\n<td>inclusionAI\/Ling-3.0-flash-Sante-fp8<\/td>\n<td>127.5B<\/td>\n<td>\u2014<\/td>\n<td>mit<\/td>\n<td><a href=\"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/2026\/11\/17\/ling-3-0-flash-sante-fp8-released\/\">Ling-3.0-flash-Sante-fp8 Text Generation Model: ~144GB Memory<\/a> (2026-11-17)<\/td>\n<\/tr>\n<tr>\n<td>deepseek-ai\/DeepSeek-V4-Pro-0813<\/td>\n<td>1650.5B<\/td>\n<td>\u2014<\/td>\n<td>mit<\/td>\n<td><a href=\"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/2026\/09\/26\/deepseek-v4-pro-0813-released\/\">DeepSeek-V4-Pro-0813 Text Generation Model: ~998GB Memory, GGUF Builds<\/a> (2026-09-26)<\/td>\n<\/tr>\n<tr>\n<td>nvidia\/DeepSeek-V4-Pro-0813-nvfp4-DSpark<\/td>\n<td>1650.5B<\/td>\n<td>\u2014<\/td>\n<td>mit<\/td>\n<td><a href=\"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/2026\/09\/10\/nvidia-releases-deepseek-v4-pro-nvfp4\/\">DeepSeek-V4-Pro-0813-nvfp4-DSpark: ~1005GB Memory<\/a> (2026-09-10)<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<\/div>\n<p><!-- \/lmw:peers --><\/p>\n<h2>How to Get It<\/h2>\n<p>Model weights are released on Hugging Face in safetensors (FP8) format. The MIT license applies, and since it is not a gated model, it can be used without submitting a special approval request on Hugging Face.<\/p>\n<p>For the inference environment, startup methods using SGLang and vLLM are officially guided. To utilize the MoE architecture and long context, operating in a multi-GPU configuration combining Tensor Parallelism (TP) and Expert Parallelism (EP) is anticipated.<\/p>\n<h3>Running with SGLang<\/h3>\n<p>An official Docker image is provided, allowing immediate use of an environment that includes speculative decoding (NEXTN) and optimizations for Hopper.<\/p>\n<p>Pulling the Docker image:<\/p>\n<pre><code class=\"language-bash\">docker pull lmsysorg\/sglang:dev-ling-3.1-flash\n<\/code><\/pre>\n<p>Example of starting an inference server using 8 GPUs (such as H200):<\/p>\n<pre><code class=\"language-bash\">docker run --gpus all --network host --ipc=host --rm \\\n  lmsysorg\/sglang:dev-ling-3.1-flash \\\n  python3 -m sglang.launch_server \\\n    --model-path inclusionAI\/Ling-3.1-flash-1019 \\\n    --tp 8 --ep 8 \\\n    --speculative-algorithm NEXTN \\\n    --host 0.0.0.0 --port 30000\n<\/code><\/pre>\n<p>For a 4-GPU GB300 environment, specify <code>--tp 4 --ep 4<\/code>. Additionally, to enable decode optimizations for the Hopper architecture, set the environment variables <code>SGLANG_ENABLE_FP8_BLOCK_GEMV=1<\/code> and <code>SGLANG_ENABLE_FUSED_KDA_OUTPUT_QUANT=1<\/code>.<\/p>\n<p>Requests can be sent to the started server in OpenAI-compatible API format using curl commands or similar. To enable the thinking function, specify <code>enable_thinking: true<\/code> within <code>chat_template_kwargs<\/code>.<\/p>\n<pre><code class=\"language-bash\">curl -s http:\/\/localhost:30000\/v1\/chat\/completions \\\n  -H &quot;Content-Type: application\/json&quot; \\\n  -d '{&quot;model&quot;: &quot;Ling-3.1-flash-1019&quot;,\n       &quot;messages&quot;: [{&quot;role&quot;: &quot;user&quot;, &quot;content&quot;: &quot;What is the capital of France?&quot;}],\n       &quot;chat_template_kwargs&quot;: {&quot;enable_thinking&quot;: true},\n       &quot;stream&quot;: true,\n       &quot;temperature&quot;: 1.0,\n       &quot;top_k&quot;: 20,\n       &quot;top_p&quot;: 0.95}'\n<\/code><\/pre>\n<h3>Running with vLLM<\/h3>\n<p>When using vLLM, build the environment using uv or a similar tool before launching.<\/p>\n<p>Example of starting an inference server:<\/p>\n<pre><code class=\"language-bash\">vllm serve &quot;$MODEL_PATH&quot; \\\n    --port &quot;$PORT&quot; \\\n    --trust-remote-code \\\n    --served-model-name auto \\\n    --tensor-parallel-size 8 \\\n    --gpu-memory-utilization 0.85 \\\n    --enable-prefix-caching \\\n    --enable-expert-parallel \\\n    --mamba-cache-mode align \\\n    --enable-auto-tool-choice \\\n    --tool-call-parser ling3 \\\n    --reasoning-parser ling3 \\\n    --speculative-config '{&quot;method&quot;:&quot;mtp&quot;,&quot;num_speculative_tokens&quot;:3}'\n<\/code><\/pre>\n<p>Example of client access:<\/p>\n<pre><code class=\"language-bash\">curl -s http:\/\/${MASTER_IP}:${PORT}\/v1\/chat\/completions \\\n  -H &quot;Content-Type: application\/json&quot; \\\n  -d '{&quot;model&quot;: &quot;auto&quot;, \n       &quot;messages&quot;: [{&quot;role&quot;: &quot;user&quot;, &quot;content&quot;: &quot;What is the capital of France?&quot;}],\n       &quot;chat_template_kwargs&quot;: {&quot;enable_thinking&quot;: true},\n       &quot;stream&quot;: true,\n       &quot;temperature&quot;: 1.0, \n       &quot;top_k&quot;: 20,\n       &quot;top_p&quot;: 0.95}'\n<\/code><\/pre>\n<p>Official guidelines recommend setting sampling parameters to <code>temperature=1.0<\/code>, <code>top_p=0.95<\/code>, and <code>top_k=20<\/code>, along with enabling the thinking function (<code>enable_thinking: true<\/code>).<\/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>2026-11-17<\/td>\n<td>bartowski<\/td>\n<td>GGUF (imatrix)<\/td>\n<td><a href=\"https:\/\/huggingface.co\/bartowski\/Ling-3.1-flash-1019-GGUF\">bartowski\/Ling-3.1-flash-1019-GGUF<\/a><\/td>\n<td>IQ1_S 146.4GB (does not fit a single consumer GPU)<\/td>\n<\/tr>\n<tr>\n<td>2026-11-17<\/td>\n<td>mlx-community<\/td>\n<td>MLX (imatrix)<\/td>\n<td><a href=\"https:\/\/huggingface.co\/mlx-community\/Ling-3.1-flash-1019-oQ4e-mtp\">mlx-community\/Ling-3.1-flash-1019-oQ4e-mtp<\/a><\/td>\n<td>MLX 4bit 368.9GB (does not fit a single consumer GPU)<\/td>\n<\/tr>\n<tr>\n<td>2026-11-17<\/td>\n<td>mlx-community<\/td>\n<td>MLX (imatrix)<\/td>\n<td><a href=\"https:\/\/huggingface.co\/mlx-community\/Ling-3.1-flash-1019-oQ2e-mtp\">mlx-community\/Ling-3.1-flash-1019-oQ2e-mtp<\/a><\/td>\n<td>MLX 2bit 215.9GB (does not fit a single consumer GPU)<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<\/div>\n<p>File sizes of each build:<\/p>\n<ul>\n<li>Available builds in bartowski\/Ling-3.1-flash-1019-GGUF: IQ1_S 122.0GB \/ IQ1_M 135.1GB \/ IQ2_XXS 159.1GB \/ IQ2_XS 165.2GB \/ IQ2_S 175.5GB \/ IQ2_M 198.4GB \/ Q2_K 204.1GB \/ IQ3_XXS 230.4GB \/ IQ3_XS 242.5GB \/ Q3_K_S 243.6GB \/ Q3_K_M 256.2GB \/ Q3_K_L 268.2GB \/ IQ3_M 276.9GB \/ IQ4_XS 295.3GB \/ Q4_0 304.2GB \/ Q4_K_S 313.5GB \/ Q4_K_M 333.4GB \/ IQ4_NL 333.4GB \/ Q4_1 335.4GB \/ Q5_K_S 378.2GB \/ Q5_K_M 406.4GB \/ Q6_K_S 445.9GB \/ Q6_K 467.1GB \/ Q8_0 566.6GB<\/li>\n<li>Available builds in mlx-community\/Ling-3.1-flash-1019-oQ4e-mtp: MLX 4bit 307.4GB<\/li>\n<li>Available builds in mlx-community\/Ling-3.1-flash-1019-oQ2e-mtp: MLX 2bit 179.9GB<\/li>\n<\/ul>\n<p>In addition, 8 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\/11\/17\/ling-31-flash-released\/\">Ling-3.1-flash Hybrid Reasoning MoE Model: ~641GB Memory, GGUF Builds<\/a><\/li>\n<li><a href=\"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/2026\/11\/17\/ling-3-0-flash-sante-fp8-released\/\">Ling-3.0-flash-Sante-fp8 Text Generation Model: ~144GB Memory<\/a><\/li>\n<li><a href=\"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/2026\/11\/17\/zoomgain-4b-fine-grained-vlm\/\">ZoomGain-4B: 4GB+ VRAM, GGUF Builds<\/a><\/li>\n<li><a href=\"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/2026\/11\/17\/zoomgain-9b-open-weight-vision-language-model\/\">ZoomGain-9B Fine-Grained Visual Language Model: 8GB+ 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 does not fit even in 80GB; its smallest build needs about 641GB) \u2192 <a href=\"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/vram-guide-en\/\">VRAM quick reference, including models over 80GB<\/a><\/li>\n<li><strong>Explore the same model family<\/strong> \u2192 <a href=\"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/model-inclusionai-ling-3-1-en\/\">Ling-3.1 family overview (3 articles, 5 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> \/ <a href=\"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/engine-vllm-en\/\">vLLM<\/a><\/li>\n<li><strong>What FP8 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-inclusionai-en\/\">inclusionAI (Ant Group): 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\/inclusionAI\/Ling-3.1-flash-1019\">inclusionAI\/Ling-3.1-flash-1019 (Hugging Face)<\/a><\/li>\n<li><a href=\"https:\/\/docs.sglang.io\/cookbook\/autoregressive\/InclusionAI\/Ling-3.1-flash\">SGLang Ling-3.1-flash cookbook<\/a><\/li>\n<li><a href=\"https:\/\/huggingface.co\/inclusionAI\/AI-Transparency\/resolve\/main\/Ling-3.1_TDS-Summary.pdf\">Ling-3.1 Training-content summary (PDF)<\/a><\/li>\n<li><a href=\"https:\/\/huggingface.co\/inclusionAI\/AI-Transparency\">inclusionAI AI-Transparency Documentation Index<\/a><\/li>\n<\/ul>\n<p><!-- lmw:updates --><\/p>\n<h2>Update History<\/h2>\n<ul>\n<li>2026-11-17: Added converted builds to \u201cQuantized and Converted Variants\u201d: bartowski\/Ling-3.1-flash-1019-GGUF, mlx-community\/Ling-3.1-flash-1019-oQ4e-mtp, mlx-community\/Ling-3.1-flash-1019-oQ2e-mtp<\/li>\n<li>2026-11-17: 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>inclusionAI has released Ling-3.1-flash-1019, a 560B hybrid linear MoE model with a 1M token context window, optimized for agent workflows.<\/p>\n","protected":false},"author":1,"featured_media":16955,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[310],"tags":[1691,3334,4151,165,169,1547,592,1555],"class_list":["post-13753","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-new-models","tag-inclusionai-en","tag-ling-3-1-flash-en","tag-ling-3-1-flash-1019-en","tag-moe-en","tag-sglang-en","tag-verified","tag-vllm-en","tag--en"],"lang":"en","translations":{"en":13753,"ja":13751},"_links":{"self":[{"href":"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/wp-json\/wp\/v2\/posts\/13753","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=13753"}],"version-history":[{"count":5,"href":"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/wp-json\/wp\/v2\/posts\/13753\/revisions"}],"predecessor-version":[{"id":17950,"href":"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/wp-json\/wp\/v2\/posts\/13753\/revisions\/17950"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/wp-json\/wp\/v2\/media\/16955"}],"wp:attachment":[{"href":"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/wp-json\/wp\/v2\/media?parent=13753"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/wp-json\/wp\/v2\/categories?post=13753"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/wp-json\/wp\/v2\/tags?post=13753"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}