Moonshot AI
Kimi K3 benchmark results
9 results on 5 benchmarks from 3 sources.
Also listed as: Kimi K3 (Max) · kimi-k3_low · Kimi K3 (low) · kimi-k3_high · Kimi K3 (high) · kimi-k3_max · Kimi K3 (max)
All benchmark results
| Benchmarks | Score | Rank | Setting | Source type | Date |
|---|---|---|---|---|---|
| LMArena Text | 1487 | 18 | Overall, style control on; 35379 votes; 95% CI 1483–1492“Kimi K3 (Max)” | Official leaderboard | Oct 22, 2026 |
| LMArena WebDev | 1655 | 16 | Overall; 20258 votes; 95% CI 1649–1661“Kimi K3 (Max)” | Official leaderboard | Oct 22, 2026 |
| GPQA Diamond | 93.1% | 20 | Epoch AI run, accuracy; effort: max | Independent run | Jul 16, 2026 |
| GPQA Diamond | 91.9% | 26 | Epoch AI run, accuracy; effort: high | Independent run | Aug 7, 2026 |
| GPQA Diamond | 84.8% | 87 | Epoch AI run, accuracy; effort: low | Independent run | Aug 7, 2026 |
| AIME (OTIS Mock AIME 2024–2025) | 97.2% | 36 | Epoch AI run, OTIS Mock AIME 2024–2025; effort: max | Independent run | Jul 16, 2026 |
| AIME (OTIS Mock AIME 2024–2025) | 93.3% | 56 | Epoch AI run, OTIS Mock AIME 2024–2025; effort: high | Independent run | Aug 7, 2026 |
| AIME (OTIS Mock AIME 2024–2025) | 68.9% | 131 | Epoch AI run, OTIS Mock AIME 2024–2025; effort: low | Independent run | Aug 7, 2026 |
| τ²-bench | 37.1% | 11 | banking_knowledge domain, pass^1; effort: max; submitted by Sierra; v1.0.1 | Official leaderboard | Jul 24, 2026 |
Where Kimi K3 ranks
Primary setting, one best-sourced score per model (official / independent first); neighbours in brackets.
| Benchmark | Score | Rank | Nearby models | #1 |
|---|---|---|---|---|
| LMArena Text | 1487 | #15 / 392 | Gemini 3.7 Flash 1490 · Muse Spark 1489 · Gemini 3.1 Pro 1487 · Gemini 3 Pro 1486 | Gemini 4 Argon1517 |
| GPQA Diamond | 93.1% | #19 / 217 | GPT-5.4 93.3% · DeepSeek V4.1 Flash 1017 93.2% · Grok 4.7 92.7% · Qwen 3.8 Max 92.7% | Gemini 4 Argon96.4% |
| LMArena WebDev | 1655 | #14 / 135 | DeepSeek V4.1 Flash 1664 · Muse Spark 1.3 1657 · Muse Wonder 1651 · Grok 4.7 1637 | Claude Opus 5.51807 |
| AIME (OTIS Mock AIME 2024–2025) | 97.2% | #33 / 200 | Grok 4.5 97.8% · Gemini 3.7 Flash 97.2% · DeepSeek-V4-Pro 96.7% · Kimi K2.6 96.1% | Gemini 4 Argon100.0% |
Other Moonshot AI models
Related on this site
FAQ
Where does Kimi K3 rank best?
LMArena Text: #15 of 392 (1487); #1 there is Gemini 4 Argon (1517).
Where do these Kimi K3 scores come from?
From Epoch AI Benchmarking Hub, LMArena leaderboard dataset, and τ²-bench leaderboard submissions. Each row links to the original leaderboard, dataset or report, and vendor-reported numbers are labelled.