{
 "id": "r7d085f93d9d",
 "found": true,
 "parent_id": "r8f4c29f3a52",
 "created_at": "2026-08-31 22:02:37.626467+00:00",
 "rows": [
  {
   "release_date": "2026",
   "release_month": "2026",
   "organization": "DeepSeek-AI",
   "model": "DeepSeek-V4",
   "parameters": "1.6T total (49B activated) for DeepSeek-V4-Pro, 284B total (13B activated) for DeepSeek-V4-Flash",
   "context_window": "1,000,000 tokens",
   "license": "MIT",
   "openness_class": "permissive open-source license",
   "significance": " Mixture-of-Experts language models designed for efficient million-token context intelligence, coding, reasoning, and agentic tasks.",
   "source_count": 1,
   "source_basis": "first-party model card or announcement",
   "source_url": "https://huggingface.co/deepseek-ai/DeepSeek-V4-Pro?show_file_info=model.safetensors.index.json"
  },
  {
   "release_date": "2026-02-02",
   "release_month": "2026-02",
   "organization": "Moonshot AI",
   "model": "Kimi K2.5",
   "parameters": "Total Parameters: 1T, Activated Parameters: 32B",
   "context_window": "256K",
   "license": "Modified MIT",
   "openness_class": "open weights / custom terms",
   "significance": "Native multimodal agentic model built through continual pretraining on mixed visual and text tokens",
   "source_count": 1,
   "source_basis": "first-party model card or announcement",
   "source_url": "https://huggingface.co/moonshotai/Kimi-K2.5/blob/refs%2Fpr%2F66/README.md"
  },
  {
   "release_date": "2026-02-13",
   "release_month": "2026-02",
   "organization": "Zhipu AI",
   "model": "GLM-5",
   "parameters": "744B (44B active)",
   "context_window": "1 million-token",
   "license": "MIT",
   "openness_class": "permissive open-source license",
   "significance": "rank #1 among open-source models on Vending Bench 2; achieves best-in-class performance among all open-source models in the world on reasoning, coding, and agentic tasks",
   "source_count": 10,
   "source_basis": "multi-source synthesis",
   "source_url": "https://z.ai/blog/glm-5"
  },
  {
   "release_date": "2026-02-16",
   "release_month": "2026-02",
   "organization": "Alibaba",
   "model": "Qwen 3.5",
   "parameters": "17B active / 397B total",
   "context_window": "256K",
   "license": "Apache 2.0",
   "openness_class": "permissive open-source license",
   "significance": "Qwen3.5 (397B total / 17B active MoE); 256K native context; 256 experts",
   "source_count": 4,
   "source_basis": "multi-source synthesis",
   "source_url": "https://jaredwatkins.com/research/local-llm-finetuning/fundamentals/base-models"
  },
  {
   "release_date": "2026-03-11",
   "release_month": "2026-03",
   "organization": "NVIDIA",
   "model": "Nemotron 3 Super",
   "parameters": "120.6B",
   "context_window": "Up to 1M",
   "license": "Open weights",
   "openness_class": "open model; verify terms",
   "significance": "36.0 on the Artificial Analysis Intelligence Index",
   "source_count": 4,
   "source_basis": "multi-source synthesis",
   "source_url": "https://tech-insider.org/ca/nvidia-nemotron-3-ultra-2026"
  },
  {
   "release_date": "2026-03-16",
   "release_month": "2026-03",
   "organization": "Mistral AI",
   "model": "Mistral Small 4",
   "parameters": "119B / 6B",
   "context_window": "256,000",
   "license": "Apache 2.0",
   "openness_class": "permissive open-source license",
   "significance": "unifies reasoning, multimodal understanding, and agentic coding in one model",
   "source_count": 11,
   "source_basis": "multi-source synthesis",
   "source_url": "https://hidekazu-konishi.com/entry/open_weights_llm_release_history_and_timeline.html"
  },
  {
   "release_date": "2026-04-07",
   "release_month": "2026-04",
   "organization": "Z.ai",
   "model": "GLM-5.1",
   "parameters": null,
   "context_window": null,
   "license": "MIT License",
   "openness_class": "permissive open-source license",
   "significance": "next-generation flagship model for agentic engineering, with significantly stronger coding capabilities than its predecessor",
   "source_count": 1,
   "source_basis": "first-party model card or announcement",
   "source_url": "https://z.ai/blog/glm-5.1"
  },
  {
   "release_date": "2026-04-12",
   "release_month": "2026-04",
   "organization": "MiniMax",
   "model": "MiniMax M2.7",
   "parameters": "229B total parameters, 10B active per token",
   "context_window": "200K tokens",
   "license": "Modified MIT",
   "openness_class": "open weights / custom terms",
   "significance": "Hits 56.22% on SWE-Bench Pro and 57.0% on Terminal Bench 2, with downloads already in the tens of thousands.",
   "source_count": 6,
   "source_basis": "multi-source synthesis",
   "source_url": "https://ai-tldr.dev/releases/minimax-m2-7-open-weights"
  },
  {
   "release_date": "2026-04-16",
   "release_month": "2026-04",
   "organization": "Alibaba",
   "model": "Qwen 3.6",
   "parameters": "35B-A3B and 27B",
   "context_window": "1M",
   "license": "Apache 2.0",
   "openness_class": "permissive open-source license",
   "significance": "building on Qwen3.5 while prioritizing stability, real-world coding utility, and direct community feedback",
   "source_count": 5,
   "source_basis": "multi-source synthesis",
   "source_url": "https://hidekazu-konishi.com/entry/open_weights_llm_release_history_and_timeline.html"
  },
  {
   "release_date": "2026-04-20",
   "release_month": "2026-04",
   "organization": "Moonshot AI",
   "model": "Kimi K2.6",
   "parameters": "1T total / 32B active (MoE)",
   "context_window": "256K",
   "license": "Modified MIT",
   "openness_class": "open weights / custom terms",
   "significance": "Native multimodal agentic model. MIT-equivalent below 100M MAU / $20M monthly revenue, attribution clause above.",
   "source_count": 7,
   "source_basis": "multi-source synthesis",
   "source_url": "https://fazm.ai/t/latest-open-source-llm-releases-2026"
  },
  {
   "release_date": "2026-04-29",
   "release_month": "2026-04",
   "organization": "Mistral",
   "model": "Mistral Medium 3.5",
   "parameters": "128B",
   "context_window": "256K",
   "license": "modified MIT license",
   "openness_class": "open weights / custom terms",
   "significance": "merges instruction-following, reasoning, and coding in a single set of weights",
   "source_count": 10,
   "source_basis": "multi-source synthesis",
   "source_url": "https://hidekazu-konishi.com/entry/open_weights_llm_release_history_and_timeline.html"
  },
  {
   "release_date": "2026-06-01",
   "release_month": "2026-06",
   "organization": "MiniMax",
   "model": "MiniMax M3",
   "parameters": "Undisclosed (MoE)",
   "context_window": "1M",
   "license": "MiniMax Community License",
   "openness_class": "open weights / custom terms",
   "significance": "First open-weight model pairing frontier coding, 1M context, and native multimodality. Vendor-reported 59.0% SWE-Bench Pro. Weights staged to Hugging Face after launch.",
   "source_count": 16,
   "source_basis": "multi-source synthesis",
   "source_url": "https://fazm.ai/t/latest-open-source-llm-releases-2026"
  },
  {
   "release_date": "2026-06-04",
   "release_month": "2026-06",
   "organization": "NVIDIA",
   "model": "Nemotron 3 Ultra",
   "parameters": "550B (55B active)",
   "context_window": "1M tokens",
   "license": "OpenMDW-1.1",
   "openness_class": "open weights / custom terms",
   "significance": "Frontier-scale general purpose reasoning and chat model optimized for complex agentic workflows, long-context reasoning, and high-stakes analytical workloads.",
   "source_count": 1,
   "source_basis": "first-party model card or announcement",
   "source_url": "https://huggingface.co/nvidia/NVIDIA-Nemotron-3-Ultra-550B-A55B-BF16"
  },
  {
   "release_date": "2026-06-13",
   "release_month": "2026-06",
   "organization": "Zhipu AI",
   "model": "GLM 5.2",
   "parameters": "744B (40B active)",
   "context_window": "1 million tokens",
   "license": "MIT",
   "openness_class": "permissive open-source license",
   "significance": "Successor to GLM-5.1 (vendor-reported 58.4% SWE-Bench Pro). Positioned for long-horizon agentic engineering.",
   "source_count": 20,
   "source_basis": "multi-source synthesis",
   "source_url": "https://fazm.ai/t/latest-open-source-llm-releases-2026"
  },
  {
   "release_date": "2026-07-02",
   "release_month": "2026-07",
   "organization": "Google DeepMind",
   "model": "Gemma 4",
   "parameters": null,
   "context_window": "256K tokens",
   "license": "Apache 2.0",
   "openness_class": "permissive open-source license",
   "significance": "multimodal, handling text and image input (with audio supported on E2B, E4B, and 12B) and generating text output",
   "source_count": 1,
   "source_basis": "first-party model card or announcement",
   "source_url": "https://huggingface.co/google/gemma-4-31B-it-qat-q4_0-unquantized"
  },
  {
   "release_date": "2026-07-15",
   "release_month": "2026-07",
   "organization": "Thinking Machines Lab",
   "model": "Inkling",
   "parameters": "975B-parameter",
   "context_window": null,
   "license": "Apache 2.0",
   "openness_class": "permissive open-source license",
   "significance": "Thinking Machines Lab's first model trained from scratch: a 975B-parameter open-weights multimodal Mixture-of-Experts with 41B active parameters and controllable thinking effort",
   "source_count": 3,
   "source_basis": "multi-source synthesis",
   "source_url": "https://best-ai.news/ai-model-releases-2026"
  },
  {
   "release_date": "2026-07-27",
   "release_month": "2026-07",
   "organization": "Moonshot AI",
   "model": "Kimi K3",
   "parameters": "2.8-trillion-parameter",
   "context_window": "1-million-token",
   "license": "Kimi K3 License",
   "openness_class": "open weights / custom terms",
   "significance": "matched or surpassed Anthropic's Fable 5 on key benchmarks at roughly half the cost, erasing the closed-source lead",
   "source_count": 28,
   "source_basis": "multi-source synthesis",
   "source_url": "https://edgen.tech/tr/news/post/kimi-k3-open-weight-release-draws-100k-downloads-erases-us-ai-lead"
  },
  {
   "release_date": "2026-08-10",
   "release_month": "2026-08",
   "organization": "Meta",
   "model": "Muse Glimmer",
   "parameters": "30-billion-parameter",
   "context_window": "128K",
   "license": "Apache 2.0",
   "openness_class": "permissive open-source license",
   "significance": "a 30-billion-parameter model distilled from Muse Spark",
   "source_count": 4,
   "source_basis": "multi-source synthesis",
   "source_url": "https://absolutedigitalpublishers.com/articles/llama-mistral-deepseek-qwen-gemma-and-olmo-what-open-actually-means"
  },
  {
   "release_date": "2026-08-14",
   "release_month": "2026-08",
   "organization": "Alibaba",
   "model": "Qwen3.8-27B",
   "parameters": "27B",
   "context_window": "262K",
   "license": "Apache 2.0",
   "openness_class": "permissive open-source license",
   "significance": "Artificial Analysis independently scores it 52 on its Intelligence Index \u2014 up from 38 for the architecturally identical Qwen3.6-27B, level with GLM-5.2 and DeepSeek V4 Flash, and ahead of every open model in the 40B\u2013150B range.",
   "source_count": 2,
   "source_basis": "multi-source synthesis",
   "source_url": "https://codersera.com/blog/open-source-llms-landscape-2026"
  },
  {
   "release_date": "2026-08-26",
   "release_month": "2026-08",
   "organization": "Z.ai",
   "model": "GLM-5.3-Flash",
   "parameters": "320-billion-parameter",
   "context_window": "above one million tokens",
   "license": "MIT license",
   "openness_class": "permissive open-source license",
   "significance": "the first natively multimodal model in the GLM-5 series: image and video input, a context window above one million tokens, a 63.4 result on the DeepSWE evaluation shown on the model page",
   "source_count": 2,
   "source_basis": "multi-source synthesis",
   "source_url": "https://startupfortune.com/zhipus-glm-53-flash-undercuts-claude-and-gpt-on-price-not-on-hardware"
  }
 ],
 "notes": []
}