Kimi K3
Jul 2026 · Open weight · Current · supported evidence
Largest open-weight model that is actually good. 97% of the lead score at a lower output price. Native multimodal.
Intelligence
59.7
AA Index
API price
$3 / $15
Input / output per 1M
Context
1.05M
36 tok/s
Capability
Intelligence Index59.7
Coding index76.2
Agentic index54.3
Arena Elo (offset)176
Use it when
- ▸Open-weight frontier
- ▸Long-horizon coding
- ▸Near-Claude quality
Skip if
- –Latency-critical UIs — it is not fast
Benchmarks
GPQA Diamond Graduate-level science questions designed so Google search is not enough. Still one of the cleanest knowledge/reasoning splits. | 93.5% |
|---|---|
Humanity’s Last Exam Expert-written questions across fields. Harder than MMLU; the current differentiator for “does this model actually know things.” | — |
MMLU-Pro Harder, less-saturated successor to MMLU. Classic MMLU is above 90% for every flagship and no longer ranks the field. | — |
SWE-bench Verified 500 human-validated GitHub issues. Score swings 5–15 points by harness — treat vendor numbers as an upper bound. | 93.4% |
SWE-bench Pro Harder, contamination-resistant coding eval. Currently the best public split between “can code” and “can maintain a repo.” | — |
Terminal-Bench 2.1 End-to-end tasks in a real terminal. Better proxy for coding agents than HumanEval, which is fully saturated. | — |
ARC-AGI-2 Abstract visual puzzles. Rewards generalization over memorization. GPT-5.6 Sol currently leads the published set. | — |
AIME 2025 American Invitational Mathematics Examination. Contest math; reasoning-mode models dominate. | — |
MMMU College-level multimodal understanding across diagrams, charts, and exam figures. | — |
Modalities
text · vision · tools
2.8T MoE / 16 of 896 experts
Reasoning mode
reasoning
Hybrid and reasoning models spend tokens thinking. That raises GPQA and agents, and also raises latency and bill.