Moonshot AI rewrites kimi-code, fully adopting Claude Code's architecture
Antigravity adds a leaner Gemini model to cut token waste on simple tasks and resets all account limits for the week.
Moonshot AI
Most Recent
Cerebras runs trillion-parameter Kimi K2.6 at 981 tokens per second
Tencent's new Chronicles-OCR benchmark reveals that top AI models like GPT-5 and Gemini fail miserably at recognizing ancient Chinese scripts, scoring near zero on oracle bone and bronze inscriptions.
Cursor's Composer 2.5 edges GPT-5.5, taps Colossus 2 supercomputer
Tencent's new Chronicles-OCR benchmark reveals that top AI models like GPT-5 and Gemini fail miserably at recognizing ancient Chinese scripts, scoring near zero on oracle bone and bronze inscriptions.
No job titles, AI interviews: Moonshot AI's radical workplace experiment
Recursive, founded by former DeepMind chief scientist Tim Rocktäschel, wants AI to run its own experiments. The startup just raised $650M at a $4.65B valuation.
Kimi raises $2B, valuation quadruples to $20B in six months
In court, OpenAI president Greg Brockman revealed that Elon Musk demanded 51% control of the startup to fund a Mars city — and stormed out when they refused.
Warp open-sources its terminal after 5 years, with OpenAI backing agent-driven contributions
After five years as a closed-source project, the AI terminal Warp is now open source. OpenAI is the founding sponsor, and AI agents will manage community contributions.
Kimi K2.6 tops OpenRouter, beats Claude 4.6 by 40% in usage
Moonshot's Kimi K2.6 became the most-used model on OpenRouter within a week of launch, consuming 1.88 trillion tokens and surpassing Claude Sonnet 4.6 by nearly 40%.
Kimi's next model could pack 2.5 trillion parameters and 1M context
Leaked specs suggest Kimi's upcoming K3 model will have over 2.5 trillion parameters and a 1M token context window, but the bottleneck isn't technology—it's compute.
US State Department Cables Accuse Chinese AI Startups of Stolen Models
The State Department has sent diplomatic cables to embassies worldwide, warning that Chinese AI companies DeepSeek, Moonshot AI (Kimi), and MiniMax are using 'distillation' to steal US AI models. Beijing calls the allegations 'groundless'.
Moonshot AI's FlashKDA speeds up Kimi Linear by up to 2.2x on Nvidia's latest
Moonshot AI open-sourced FlashKDA, a tool that makes its Kimi Linear model run 1.7 to 2.2 times faster on Nvidia's latest Hopper GPUs by optimizing its KDA attention mechanism.