Updated daily at 05:00 UTC
- Released
- 6 Nov 2025
- Status
- Legacy
- Context
- 262.1K
- Max output
- —
- Input per 1M tokens
- —
- Output per 1M tokens
- —
- Open weights
- Yes
- License
- modified-mit
- Parameters
- 1000B, 32B active
- Knowledge cutoff
- —
- Reasoning
- Yes
- Modalities
- text → text
Category scores
Out of 100, from independent and third-party results only, by method category-2026-09-29.1.
| Category | Score | Rank | Coverage |
|---|---|---|---|
| Knowledge & reasoning | 15.8 | Not ranked | 1 of 4 |
| Human preference | 46.6 | Not ranked | 2 of 5 |
Benchmark results
Every result we hold for this model, with who ran it and where it was published.
Knowledge & reasoning
| Benchmark | Metric | Value | Provenance | Setting | Evaluated | Source |
|---|---|---|---|---|---|---|
| Chess Puzzles | Puzzles solved | 20.0% | Independent | — | 10 Dec 2025 | Epoch AI |
| Humanity's Last Exam (HLE) | Accuracy | 44.9% | Self-reported | thinking | — | Model Card via Hugging Face Community Evals |
Scientific reasoning
| Benchmark | Metric | Value | Provenance | Setting | Evaluated | Source |
|---|---|---|---|---|---|---|
| GPQA Diamond | Accuracy | 84.2% | Independent | — | 11 Nov 2025 | Epoch AI |
| GPQA Diamond | Accuracy | 84.5% | Self-reported | thinking | — | Model Card via Hugging Face Community Evals |
Mathematics
| Benchmark | Metric | Value | Provenance | Setting | Evaluated | Source |
|---|---|---|---|---|---|---|
| FrontierMath Tier 4 (v1) | Accuracy | 0.0% | Independent | — | 4 Dec 2025 | Epoch AI |
| FrontierMath Tiers 1-3 (v1) | Accuracy | 21.4% | Independent | — | 5 Dec 2025 | Epoch AI |
| OTIS Mock AIME 2024-2025 | Accuracy | 83.1% | Independent | — | 13 Nov 2025 | Epoch AI |
Coding & software engineering
| Benchmark | Metric | Value | Provenance | Setting | Evaluated | Source |
|---|---|---|---|---|---|---|
| METR Time Horizons | Average task success (Epoch's column) | 59.2% | Third party | — | — | METR - Measuring AI Ability to Complete Long Tasks via Epoch AI |
| Terminal-Bench 2.0 | Accuracy (mean over trials) | 35.7% | Third party | — | — | https://www.tbench.ai/leaderboard/terminal-bench/2.0 via Epoch AI |
| Terminal-Bench 2.0 | Accuracy (mean over trials) | 35.7% | Third party | — | — | Terminal-Bench v2 Leaderboard via Epoch AI |
| WeirdML (v2) | Average accuracy | 42.8% | Third party | — | — | https://htihle.github.io/weirdml.html via Epoch AI |
Human preference
| Benchmark | Metric | Value | Provenance | Setting | Evaluated | Source |
|---|---|---|---|---|---|---|
| Code Arena: WebDev (arena.ai) | Arena score (Elo scale) | 1322 | Independent | — | 25 Sep 2026 | Arena |
| Code Arena: WebDev (arena.ai) | Votes | 13100 | Independent | — | 25 Sep 2026 | Arena |
| Text Arena (arena.ai) | Arena score (Elo scale) | 1430 | Independent | — | 25 Sep 2026 | Arena |
| Text Arena (arena.ai) | Votes | 63241 | Independent | — | 25 Sep 2026 | Arena |
Other Kimi K versions
- Kimi K316 Jul 2026
- Kimi K2.7 Code12 Jun 2026
- Kimi K2.620 Apr 2026
- Kimi K2.527 Jan 2026
- Kimi K211 Jul 2025
- Kimi Thinking PreviewMay 2025
- kimi-latestFeb 2025