Kimi K2.6 Targets Claude With Open-Weight Model at One-Sixth the Price
Kimi K2.6 is being presented as a lower-cost alternative to Anthropic's Claude, with chat, coding and agent tools offered at one-sixth of the price. The open-weight 1-trillion-parameter mixture-of-experts model activates 32 billion parameters per token and includes a 262,000-token context window across four chat modes: Instant, Thinking, Agent and Agent Swarm.
The package also includes Kimi Code, an open-source command-line coding agent, and Kimi Agent, which is described as generating websites, presentation decks, spreadsheets and documents. K2.6 is said to support up to 300 parallel sub-agents, sustain more than 4,000 tool calls in a single run and keep sessions going for up to 13 hours; on SWE-Bench Pro, it was listed at 58.6, ahead of GPT-5.4 xhigh at 57.7, Gemini 3.1 Pro at 54.2 and Claude Opus 4.6 at 53.4. The model is also described as self-hostable on four H100s in INT4.
From the sources (1 posts)
@_avichawlaAnthropic's in trouble, again. The entire Claude experience is now available at 1/6th the price. Kimi now does everything Claude does, powered by K2.6, a 1-trillion-parameter MoE model that activates only 32B parameters per token. It cov