Moonshot
Kimi K3
newRank 6 of 45 · release 2026-06-25 · $3 in / $15 out per MTok
Global average
79.2
first appearance
Signal breakdown
Category scores
Reasoning
90.7
Coding
81.5
Agentic Coding
62.2
Mathematics
84.4
Data Analysis
78.7
Language
85.5
Instruction Following
71.4
Trend
Score history
Analyst annotations
Commentary
Sign in to publish an annotation under your analyst identity.
No annotations yet
Be the first analyst on record for Kimi K3.
Detail
Per-task scores
| Task | Category | Score |
|---|---|---|
| zebra_puzzle | 100.0 | |
| spatial | 100.0 | |
| connections | 100.0 | |
| tablereformat | 98.0 | |
| AMPS_Hard | 97.0 | |
| math_comp | 95.1 | |
| olympiad | 91.7 | |
| consecutive_events | 89.8 | |
| typos | 84.0 | |
| theory_of_mind | 82.7 | |
| code_completion | 82.6 | |
| code_generation | 80.3 | |
| logic_with_navigation | 80.0 | |
| story_generation | 75.4 | |
| paraphrase | 74.4 | |
| plot_unscrambling | 72.6 | |
| summarize | 69.3 | |
| javascript | 68.2 | |
| simplify | 66.4 | |
| python | 65.0 | |
| integrals_with_game | 54.0 | |
| typescript | 53.3 | |
| tablejoin | 48.4 |