Alibaba-backed Chinese AI startup Moonshot unveiled Kimi K3, a 2.8-trillion-parameter open-weight model that outperforms Anthropic's Claude Fable 5 and OpenAI's GPT-5.6 Sol on some benchmarks, with Ar…
Anthropic has repeatedly extended the deadline for removing Claude Fable 5 from paid subscriptions, now set for July 19, in what critics call scarcity marketing. Meanwhile, Moonshot AI's open-weights …
Moonshot AI's open-weight Kimi K3 model scored 1,679 points to top Arena.ai's Frontend Code Arena leaderboard on July 16, beating Anthropic's Claude Fable 5 (1,631) and OpenAI's GPT-5.6 Sol (1,618). T…
Moonshot AI released Kimi K3, a new large language model that achieves state-of-the-art results on the Pelican benchmark, outperforming GPT-4o and Claude 3.5 Sonnet. The Pelican benchmark, designed to…
Kimi K3, a reasoning model released on July 16, 2026, by Kimi, scores 57 on the Artificial Analysis Intelligence Index, well above the average of 30 among comparable models, but is slower than average…
Kimi K3 ranks third on Artificial Analysis's intelligence index, trailing the leader by only two points. The independent evaluation platform assesses leading AI models across intelligence, cost, speed…
Kimi released Kimi K3, a 2.8-trillion-parameter open model with native vision and a 1-million-token context window, claiming it is the world's first open 3T-class model. While trailing proprietary mod…
Large open-weight mixture-of-experts (MoE) models are converging on lower routed-expert activation ratios, with the average ratio falling from roughly 3.7% in 2025 to 2.1% in 2026 across six major mod…
Meta, Kat Coder (KwaiPilot), and Moonshot AI released new AI models this week, all available on the Kilo platform. Kat Coder launched Pro v2.5 and Air v2.5, with Pro free for a week, featuring a new a…
Moonshot AI is preparing to release Kimi K3, an open-weight model with 2 trillion to 3 trillion parameters, directly challenging Anthropic's Claude Opus 4.8 pricing. The Beijing startup, valued at $20…
A new open letter calls for AI regulation, following Demis Hassabis's regulatory call. Twenty-six Meta employees filed a novel lawsuit alleging AI-powered software disproportionately targeted disabled…
Kimi released Kimi K3, its most capable model with 2.8 trillion parameters, built on Kimi Delta Attention and Attention Residuals, offering native visual understanding and a 1M-token context window. I…
Moonshot AI has launched Kimi K3, its most capable flagship model to date, with strong performance in coding, games/3D, and knowledge tasks. The model supports up to 1M context window and reasoning ef…
Moonshot AI launched Kimi K3, a new AI model with a 1 million token context window and Agent Swarm technology coordinating up to 300 sub-agents, quadrupling its predecessor's capacity. The model, expe…
Chinese AI lab Moonshot AI's upcoming Kimi K3 model, with 2-3 trillion parameters, is expected to match or surpass Anthropic's Opus 4.8, according to the Financial Times. The open-weight model, the la…
Moonshot AI is preparing to launch Kimi K3, the successor to its K2 model family, with beta users spotting the model in selectors and a mystery checkpoint called Kivine appearing on LM Arena. The Mixt…
Moonshot AI's leaked Kimi K3 promo page reveals a 2.5-trillion-parameter Mixture-of-Experts model with a 1 million token context window, targeting long-horizon coding and agentic workloads, but the sp…
Moonshot AI, a Beijing-based startup founded in March 2023 by Tsinghua University alumni, plans to launch its Kimi K3 model around July 15 to challenge Anthropic's Claude Opus 4.8. The model features …
China's Kimi K3 is expected to outperform the US-based Claude Opus 4.8, signaling a major shift in AI dominance. Sources confirm the Kimi K3's performance metrics could eclipse those of the Claude Opu…
Linus Torvalds defended the use of AI tools in Linux kernel development, responding to a debate about integrating the agentic LLM code review tool Sashiko into the kernel's Patchwork tracking system. …