Artificial intelligence, professionally covered
Open models are closing the gap with frontier labs. This hub follows open-weight releases, licensing fights, community fine-tunes and the infrastructure that keeps open AI competitive.
WorkSwarm puts the human inside the agent team rather than in front of it. The performance claims are the vendor's own, and there is no new model here.
2d ago
Two days after DeepSeek released Harness as a developer preview, China's largest chat platform made it a plugin — npm install, launch, scan a QR code.
2d ago
HEIR converts pre-trained models to run inference directly on homomorphically encrypted inputs. Four reference applications ship with it, and none of them are a hosted product.
2d ago
Qwen3.8-27B is dense, not mixture-of-experts, carries a 262,144-token native context, and arrives with every benchmark self-reported.
2d ago
GLM-5.3 sits on the same 753B foundation as GLM-5.2. Every claimed gain comes from post-training — and the weights are not out for two more weeks.
4d ago
New research reconstructs the March LiteLLM compromise: 2,500+ organisations potentially exposed through a file that ran whether or not you imported the library.
5d ago
Ollama, GPT4All and Msty found inside Kimsuky infrastructure, with a retrieval index pointed at stolen documents. Nothing reaches a vendor's abuse team.
6d ago
Muse Glimmer runs on one consumer GPU and carries a genuinely permissive licence. The release went out through Hugging Face and a Zuckerberg essay.
6d ago
Release candidates are supposed to shrink as a cycle ends. This one did not, and the kernel's maintainer says the bugs are being surfaced by machines.
Aug 10, 2026
The next open-weight release is set to carry a revenue-share licence for large commercial users, following Moonshot's Kimi K3 terms. Alibaba's own percentage is not set.
Aug 9, 2026
Administration officials told Meta, Anthropic, Google and OpenAI this week that the voluntary pre-release security review covers proprietary frontier systems only — a decision Anthropic has spent a year arguing against.
Aug 6, 2026
The company that started the AI price war told developers on 6 August that its API is about to get substantially more expensive. It published no rate card, no effective date and no explanation.
Aug 6, 2026
Five teams ratified a policy for rust-lang/rust on 5 August. Models may analyse, distill, check, suggest and review; generated documentation, diagnostics and soundness-critical changes are out, and any LLM text must be disclosed.
Aug 6, 2026
Qwen3.8-Max went live Monday morning with a benchmark table showing wins over GPT-5.6 Sol and Claude Fable 5. The same table, read in full, shows it losing every hard agentic-engineering row.
Aug 3, 2026
The weights landed on Hugging Face early Monday, three days after the model was announced. The licence defines an "Applicable Territory" that excludes four of the largest markets — and the modules that produce the advertised 2K output never left the API.
Aug 3, 2026
Nscale is acquiring Anyscale, folding the orchestration layer into its own cloud. The price everyone is printing was never disclosed by either party.
Jul 31, 2026
Huawei announced day-zero Ascend support for Kimi K3 — training on Atlas SuperPoD, inference through vLLM Ascend and SGLang. No throughput figures were published.
Jul 28, 2026
Dario Amodei published a personal rebuttal hours after the largest open model in history shipped. The denial is narrow, and the policy he does want would land hardest on open releases.
Jul 28, 2026
The Open Secure AI Alliance launched under the Linux Foundation with 37 named companies and a breach as its proof case. OpenAI, Anthropic, Google and Meta are all missing from the list.
Jul 28, 2026
Moonshot AI put 1.56 terabytes of Kimi K3 weights on Hugging Face on Monday afternoon. Almost every write-up called the licence a modified MIT. The file in the repository says otherwise.
Jul 28, 2026
The lab verbally told some would-be backers it will not sign in the coming days, shelving a raise of at least 10 billion yuan at a pre-money valuation of at least 480 billion — after leaked remarks attributed to its founder went viral.
Jul 26, 2026
Jensen Huang’s first post on X carried 25 signatures. Within a day the roster reached 50, with OpenAI, Google, AMD, Cisco, Cloudflare, GitHub, Block and Ollama added. Two of the biggest names in AI are still missing.
Jul 26, 2026
Hugging Face, Meta, Microsoft, Mistral, Nvidia and a16z signed. OpenAI, Anthropic and Google DeepMind did not — the clearest public split yet between the closed labs and the open-weight camp.
Jul 25, 2026
FLUX 3 generates 20 seconds of video with native audio in one pass. FLUX-mimic, released alongside it, controls robot arms — and is already in deployment testing at Audi.
Jul 24, 2026
With Y Combinator, Proton and Replit behind it, the new coalition warns that cutting off Kimi K3 and Qwen would crush startups' costs and entrench the biggest US labs — on the very day the White House accused Moonshot of IP theft.
Jul 23, 2026
China's Xiaohongshu says its beta model dots-note-3.0 solved all six IMO problems under official grading, matching just seven human contestants — with Huawei's Celia sharing the historic milestone.
Jul 23, 2026
A 1-billion-parameter open-weight model beats a 753-billion-parameter Chinese model and Gemini 3 Pro on Cisco's vulnerability-localization benchmark, running the whole suite in about 13 minutes on one H100.
Jul 22, 2026
Nvidia's CEO told Axios that American companies should absolutely be allowed to run Chinese open-weight models, dismissing the idea that a downloaded model is a backdoor to Beijing.
Jul 22, 2026
A 30-person startup spun off from AMD-backed Moreh trained Motif-3-Beta on about 700 government GPUs — and became the highest-ranked model builder outside the US and China.
Jul 21, 2026
Days after Kimi K3 launched, Chinese lab Moonshot stopped selling new subscriptions, saying demand had pushed its GPU clusters to the limit — and split its plans into two tiers to ration compute.
Jul 20, 2026
