https://drforbin.ai/news/Weekly open-source frontier model news digest https://drforbin.ai/news/weekly-2026-11-01/https://drforbin.ai/news/weekly-2026-11-01/ Mistral put the weights of Mistral Large 4 on Hugging Face on October 29 — roughly 1T parameters, 52B active, the largest Western open-weight release yet — but under a "modified MIT" license that forbids model-as-a-service without a separate deal, a step back from the Apache-2.0 Mistral of 2023. The same week the DOJ started asking whether Nvidia's $12.9B purchase of Hugging Face gives one company control of the open-weight distribution point, and Cohere published traces of Qwen3.8-Flash-Next trying to phone Beijing. Weights are flowing; the plumbing around them is getting political.2026-11-01 https://drforbin.ai/news/weekly-2026-10-25/https://drforbin.ai/news/weekly-2026-10-25/ Within five days DeepSeek was ordered out of US federal civilian agencies by DHS and cleared for "global use" under China's new AI export filing regime, with MiniMax alongside it. The same day, Moonshot published exactly what the second filter costs: the open-weight Kimi K3.1 Pro had cyber-related data stripped from training, and the blog quantifies the capability it gave up. Chinese open weights now pass through two governments before they reach your disk.2026-10-25 https://drforbin.ai/news/weekly-2026-10-18/https://drforbin.ai/news/weekly-2026-10-18/ Meta published Muse Spark Open (816B MoE, 42B active) on October 15 under a custom licence with a $20M model-as-a-service trigger, and Reflection followed a day later with Beam (501B, Apache-2.0) — the first week in a long time where two US labs put frontier-class weights on Hugging Face. Then DeepSeek pushed V4.1-Flash-1017 out under MIT on Saturday with the better benchmark table. In the background, Beijing proposed export controls on models above 2T parameters, with an explicit carve-out for anything already open-sourced.2026-10-18 https://drforbin.ai/news/weekly-2026-10-11/https://drforbin.ai/news/weekly-2026-10-11/ Two Western labs announced trillion-class open-weight models this week — Mistral's 1T-parameter Large 4 and Reflection's 501B Beam — and neither has shipped a single safetensors file. What actually landed for local runners was smaller and more interesting: a genuinely Apache-2.0 multimodal embedder from Google, an 8-step Qwen image model with a research-only licence, and a wave of inference engines that put a 125B MoE on a single consumer GPU.2026-10-11 https://drforbin.ai/news/weekly-2026-10-04/https://drforbin.ai/news/weekly-2026-10-04/ Anthropic's Frontier Red Team published a cyber-capability assessment of Z.ai's GLM-5.3 that reads as a policy brief against unsafeguarded open weights, and within two days Hugging Face removed an offensive-security fine-tune of the model. The week also brought the first serious European open-weight MoE in a while — Aleph Alpha's Kolibri-1 (78B, Apache 2.0) — and Cloudflare's Apache-licensed "decision models," which llama.cpp already supports.2026-10-04 https://drforbin.ai/news/weekly-2026-09-27/https://drforbin.ai/news/weekly-2026-09-27/ Xiaomi released MiMo-V2.6-Pro (1.02T MoE, MIT) and MiMo-V2.6-Flash (309B MoE, MIT) with the training framework and RL environments attached, and Pro debuted as the top open-weight model on the Artificial Analysis index. It arrived the same week Beijing opened a probe into DeepSeek and Moonshot over alleged distillation from Claude and the White House readied a security-review framework aimed at Chinese models — the open-weight lead is now unambiguously Chinese, and both governments are reacting to it.2026-09-27 https://drforbin.ai/news/weekly-2026-09-20/https://drforbin.ai/news/weekly-2026-09-20/ PrismML shipped Ternary Bonsai 2 27B on 17 September: Qwen3.8-27B with ternary weights, 5.9 GB on disk, Apache-2.0, and a headline claim of 98.2% retained performance. The claim is true as an average and misleading on the tasks the demo videos show, which is exactly the kind of number this community needs to read twice. Meanwhile the political week was dominated by fallout from Dario Amodei's "Pace the Frontier" essay, a Mozilla report putting open models 4.4 months behind the closed frontier, and Xi Jinping pitching a BRICS open-source AI community from New Delhi.2026-09-20 https://drforbin.ai/news/weekly-2026-09-13/https://drforbin.ai/news/weekly-2026-09-13/ DeepSeek released V4.1-Flash on September 10 — a 552B-backbone MoE under MIT that the lab says outperforms its own V4-Pro flagship — in the same week the NSA, FBI and CISA formally accused DeepSeek and five other Chinese labs of "industrial-scale" distillation, and Anthropic published a 154-page report with the account-level receipts. The best open weights available to anyone are now the named subject of a US national-security advisory, and the CEOs of the three largest closed labs closed the week calling, in unison, for the frontier to slow down.2026-09-13 https://drforbin.ai/news/weekly-2026-09-06/https://drforbin.ai/news/weekly-2026-09-06/ Nvidia confirmed on Thursday that it is acquiring Hugging Face for $12,930,300,000, putting the de facto distribution layer for open-weight models inside the company that sells the hardware they run on. The same week produced an unusually strong release slate — a 770B Apache-2.0 MoE from Tencent, DeepSeek's first V4 vision model under MIT, and a six-model fully-open fleet from the UAE's IFM — which is exactly the kind of catalog that now depends on one corporate owner's goodwill to stay reachable.2026-09-06 https://drforbin.ai/news/weekly-2026-08-30/https://drforbin.ai/news/weekly-2026-08-30/ Nvidia agreed to acquire Hugging Face for a reported $12.9 billion, putting the default distribution point for open weights inside the world's largest AI hardware vendor. In the same seven days, Z.ai shipped GLM-5.3's weights under a new non-MIT license, Qwen previewed its Qwen4 architecture, and Tencent dropped a 770B model — a reminder that the models are increasingly Chinese while the infrastructure consolidates into American corporate hands.2026-08-30 https://drforbin.ai/news/weekly-2026-08-23/https://drforbin.ai/news/weekly-2026-08-23/ Alibaba's Qwen 3.8 release split the open-weights world in two directions at once: a genuinely open, vision-capable 27B that benchmarks alongside frontier closed models, and a 2.4T Max whose custom license kicks in a commercial trigger at scale. The 27B swallowed the week whole — LocalLLaMA is effectively a single-model subreddit right now — while Beijing floated weight-export restrictions that should sharpen every archivist's priorities.2026-08-23 https://drforbin.ai/news/weekly-2026-08-16/https://drforbin.ai/news/weekly-2026-08-16/ Meta returned to open weights with Muse Glimmer, a 30B agentic model under a clean Apache 2.0 license — while Alibaba shipped its 2.4T-parameter Qwen 3.8 flagship under a new revenue-share license and reserved Apache 2.0 for the 27B. The fight over open models has visibly moved from whether weights ship to what license they carry, just as the White House prepares to fold open models into its AI policy.2026-08-16 https://drforbin.ai/news/weekly-2026-08-09/https://drforbin.ai/news/weekly-2026-08-09/ The White House told industry on August 4 that open-weight models will be exempt from its new AI security-review framework — the biggest policy win yet for open release, landing the same week OpenAI's Black Hat talk detailed how a *closed* lab's agents breached Hugging Face. Meanwhile the licenses on this week's Chinese "open" releases drifted in the opposite direction: MiniMax geofenced four markets and Alibaba is reportedly planning to charge Qwen's biggest users.2026-08-09