Lilian Weng, co-founder of Thinking Machines Lab, posted a cryptic message on July 27 calling her decision 'hard and sad,' without specifying its nature, following the July 15 release of the company's…
Celeris launched Celeris-1 on July 27th, a diffusion-based language model API that completes short reasoning calls in milliseconds and generates long responses at 1,664 tokens per second, targeting vo…
Enigma emerged from stealth on July 27th with a $71 million seed round led by Index Ventures and Ribbit Capital, and opened 100 real robots to browser control via Robots.online. The company, founded b…
Waddle Labs launched an early-access API on July 27th that uses language-model agents to write control programs for robots, claiming it can produce a working policy from a natural-language prompt in 2…
Nvidia CEO Jensen Huang said in a video posted to X on July 27 that Anthropic's Mythos language model 'should absolutely be available to everyone' after an interviewer claimed the model can 'break int…
Feyn Labs released FeyNoBg, a background-removal model, and its open-source NoBg training library on July 21, claiming the model achieved the highest published S-measure on four of eight benchmarks an…
Microsoft on July 27, 2026 integrated MAI-Cyber-1-Flash into its multi-agent security platform MDASH, claiming the model achieves a 96% score on the CyberGym benchmark—12 points higher than Mythos—at …
Atlas Discovery launched out of stealth with foundation models that predict individual patient drug responses in clinical trials, targeting the costly gap between preclinical results and human efficac…
Synopsys on July 27 announced that its autonomous debugging and physical implementation workflows for chip design are available for evaluation through Microsoft Discovery, with AMD testing their use o…
Kimi.ai released the model weights and a technical report for K3, a 2.8-trillion-parameter mixture-of-experts (MoE) model with native visual understanding and a 1 million-token context window, on July…
Wiz's Project Atlas, a vulnerability research system using specialized AI agents, scored 90.9% on the CyberGym benchmark, surpassing published results from Anthropic, OpenAI and Microsoft, according t…
Nvidia has made a substantial investment in Safe Superintelligence (SSI), the AI research lab founded by Ilya Sutskever, and will provide SSI with access to its Vera Rubin systems, increasing SSI's co…
Former Anthropic engineer Adi Baradwaj alleged in a July 26, 2026 X thread that the AI startup lowered safety safeguards on its Claude models in exchange for large committed-spend contracts. Baradwaj …
Nvidia is discussing a roughly $250 billion financial backstop to help OpenAI lease a 10-gigawatt data center project in southern Ohio, according to a Reuters report citing The Wall Street Journal. Th…
An X user leaked that Z.ai's upcoming GLM-5.5 large language model will beat Fable 5 and compete with Mythos, with a launch target of August 2026. The leak suggests Z.ai may skip the GLM-5.3 version e…
On July 26, 2026, Om Patel warned on X that artifacts generated by Anthropic's Claude large language model and published to the web are being indexed by Google search, exposing internal company dashbo…
Soar AI on July 26, 2026, released what it calls the first Model Context Protocol (MCP) interface for booking flights at market rates with no extra fees, supporting x402 payments and compatibility wit…
Invideo's Agent One ranked first in Physion Labs' independent benchmark of AI video agents, released in mid-July 2026, after the company requested inclusion. Agent One scored #1 on twelve of sixteen m…
An unverified social media post from July 26, 2026, claims Meta will launch an AI harness and open-source models, according to X user RK citing Alexandr Wang. The claim, which has not been confirmed b…
On July 26, 2026, a user known as acidvegas launched stolencompute.com, a site that aggregates publicly exposed AI model endpoints for free inference, including Kimi 1T and Deepseek 765B. The service …