Kimi — the most intelligent open weights in the world.
Moonshot AI builds the Kimi family. Kimi K3 is the most intelligent open-weight model released to date — #4 overall on aggregate intelligence at launch — and Kimi K2.7 Code is a dedicated coding model that rivals models many times its price.
Rankings refresh dailyScored on 6 criteriaNo paid rankings
Kimi K3 was #4 globally on aggregate intelligence at launch — the top open-weight result
Kimi K2.7 Code: dedicated agentic coding at $0.95/1M input
2.8T-parameter flagship with a 1M token context window
2 models
All Moonshot AI Models
Every Moonshot AI model in the directory, ranked by overall capability score.
MoonshotPremium
Kimi K3
Moonshot's 2.8-trillion-parameter multimodal reasoning flagship with always-on thinking — the largest open-weight model ever released and the closest Chinese challenger to the Western frontier.
Verdict
Closest Chinese challenger to the frontier — #4 overall on intelligence.
Quality score
93%
Pricing
$3.00/1M in
$15.00/1M out
Speed
Deliberate
2/5 speed
Context
1M tokens
Released July 16, 2026; open weights July 26. Cache-hit input $0.30/1M. Subscriptions: Adagio (free) to Vivace $199/mo; full 1M context only on Allegro ($99) and up. New signups paused July 19 near GPU capacity, reopening in batches.
An open-weight 1T-parameter MoE (32B active) coding specialist tuned for long-horizon agentic software engineering with markedly better token efficiency than its predecessor.
Verdict
Value coding specialist — 1T MoE agentic coder at budget prices.
Quality score
71%
Pricing
$0.95/1M in
$4.00/1M out
Speed
Fast
4/5 speed
Context
256k tokens
Model ID kimi-k2.7-code; weights on Hugging Face June 12, 2026. Kimi Code membership from $19/mo.
Pricing changes, new releases, and ranking shifts — straight to your inbox.
No spam. Useful updates only. Affiliate disclosures always clearly labeled.
Moonshot AI FAQ
What is Moonshot's best AI model in 2026?
Kimi K3 ($3/$15) — the most intelligent open-weight model released to date, #4 overall on aggregate intelligence at launch with a 1M context. For coding specifically, Kimi K2.7 Code ($0.95/$4) is the better value: strong agentic coding at a fraction of K3's price.
Can I self-host Kimi models?
The weights are open, but Kimi K3 is 2.8T parameters — impractical to self-host for almost everyone. In practice you access Kimi via the first-party API (platform.moonshot.ai), kimi.com for consumer use, or the Kimi Code subscription ($19/month).
How does Kimi compare to Claude and GPT?
Kimi K3 ranked #4 on aggregate intelligence at launch — above most closed models, below Claude Opus 5 and GPT-5.6 Sol. It is the strongest answer to 'what is the smartest model with open weights?' For pure coding, Claude leads on SWE-Bench Pro and K2.7 Code competes on value rather than ceiling.