According to TheRundownAI, OpenAI paused Astra safety rollout, Loom pairs with ChatGPT for onboarding, and Kimi K3 faces jailbreak testing.

According to @godofprompt, Kimi K3 delivered richer textures and smoother motion than Claude Fable 5 when rebuilding a 3D globe dashboard from one prompt.

According to Ethan Mollick, Qwen 3.8 Max is solid but lags Kimi K3 on shader generation quality based on hands-on tests.

According to The Rundown AI, Moonshot released Kimi K3 weights as Anthropic clarified open weights, while Microsoft unveiled a new cybersecurity model.

According to @emollick, Kimi K3 releases open weights, a 2.8T MoE with 1M context and native vision, plus kernels and MoE libraries for scale.

According to KyeGomezB, Kimi K3 debuts a 2.8T MoE with 1M tokens, native vision, Attention Residuals, Kimi Delta Attention, MLA, and multi-stage RL.

According to TheRundownAI, Moonshot published Kimi K3 weights, a 2.8T MoE with vision and 1M tokens, plus kernels and MoE libs for scalable agents.

According to AINewsOfficial_ on X, Moonshot Kimi K3 challenges Google Gemini Flash as new robot hands from Generative Bionics and Mimic debut.

According to emollick, the US may ban Chinese open-source models, raising security and industrial policy questions amid booming investments.

According to @emollick, Kimi K3 produced a 32 page chain of thought with loops and dead ends when asked to pick two poems.

According to emollick, Kimi K3 used 95.5% English in its chain-of-thought for a Chinese prompt, signaling training bias and evaluation gaps.

According to Ethan Mollick, the UK AI Security Institute will benchmark Kimi K3 soon; current results show GLM-5.2 matches Opus 4.5 while V4-Pro trails Sonnet 4.5.

According to TheRundownAI, Kimi K3 advances open models, Google delays AI upgrade, GPT Live speeds travel plans, plus 4 new tools and key skills.

According to @CNBC, Moonshot AI says Kimi K3 beats top OpenAI and Anthropic models on select benchmarks while still trailing overall leaders.

According to emollick, Kimi K3 leads Arena frontend ELO, but user-judged scores are limited for benchmarking and can favor prompt-tuned chat UX.

According to soumithchintala, Kimi K3 ships 2.8T params, 1M context, multimodal, faster decoding, and will open weights by July 27, 2026, per Kimi.ai.

According to KyeGomezB, Kimi K3 now leads Code Arena over Claude Fable 5 and GPT 5.6 Sol, signaling competitive pressure on US open source models.

According to @emollick, a one-shot procedural harbor town benchmark now compares GPT-5.6 Pro, Fable, Kimi K3, and Inkling.

According to TheRundownAI, Kimi K3 tops Claude Opus 4.8 and rivals GPT 5.6 Sol and Fable at Sonnet pricing, signaling a strong open source push.

According to @emollick, Kimi K3 generated a twigl shader of neo‑gothic ocean towers and improved it on prompt, showing strong open weights performance.