Kimi K2
Trillion-parameter open MoE built for agentic work.
About
Kimi K2 is Moonshot AI's 1-trillion-parameter mixture-of-experts model with 32B active parameters per token, trained on 15.5T tokens with the MuonClip optimizer. Released July 2025 under a modified MIT license, it was the largest open-weights model of its time and debuted as the top open model on the LMSYS Chatbot Arena.
It is explicitly built for agentic coding and tool use rather than chat alone. The license is permissive but not OSI-open — very large commercial deployments must display Kimi K2 attribution — and serving a 1T model demands serious GPU infrastructure.
Specs
Context window
Pricing
Free quota
API access
Weights
Released
Benchmarks
Top open-weights model on LMSYS Chatbot Arena at launch (user-voted, July 2025)
Provider-reported unless a source says otherwise. Benchmarks are marketing-adjacent by nature — treat them as directional.
Use cases
Trust score — 9/10
Scored 9/10 (verified tier, metrics-v1). Evidence: Massive independent coverage (Wikipedia, Simon Willison, Taskade); LMSYS Arena debut documented; Arena ranking is user-voted, not a formal benchmark; Manual review: landmark open-weights release; license is modified-MIT, not OSI-open. License: Modified MIT — attribution ('Kimi K2' displayed) required for products with 100M+ MAU or $20M+/mo revenue; not OSI-compliant open source.
How the trust score is built
Full methodology →- Adoption · stars/downloads, log-scaled vs peers
- 9.0/10
- Maintenance · recency, releases, issue load
- 9.0/10
- Openness · license + docs
- 8.0/10
- Manual review · sanity read + red-flag check
- 9.0/10
- · Massive independent coverage (Wikipedia, Simon Willison, Taskade); LMSYS Arena debut documented
- · Arena ranking is user-voted, not a formal benchmark
- · Manual review: landmark open-weights release; license is modified-MIT, not OSI-open