Kimi K2 Thinking
| Released | 2025-11-06 |
|---|---|
| Price per 1M | $0.60 in · $2.50 out |
| Overall score | 48.9 · rank 124 |
| Context window | 262,144 tokens |
| Provider | Moonshot AI |
| Weights | Open weights |
Scores
| Dimension | Score | Rank | Benchmarks |
|---|---|---|---|
| Overall: Balanced | 48.9 | 124 | 12 |
| Overall: Coding | 48.0 | 120 | 7 |
| Overall: Reasoning | 48.7 | 127 | 6 |
| Overall: Agentic work | 47.9 | 130 | 7 |
| Overall: Frontend | 48.0 | 112 | 2 |
| Coding | 48.0 | 132 | 2 |
| Agents | 47.6 | 124 | 2 |
| Reasoning | 49.2 | 122 | 3 |
| Math | 50.1 | 112 | 2 |
| Knowledge | 44.8 | 117 | 1 |
| General | 56.1 | 57 | 2 |
| Value vs price | +0.9 | 113 |
Evidence
Coding 48.0 ±3.3
| Benchmark | Result | Source | Date |
|---|---|---|---|
| SWE-bench Verified (Vals run) | 60.2% | Vals AI | 2026-09-01 |
| LiveCodeBench (Vals run) | 63.1% | Vals AI | 2026-09-01 |
Agents 47.6 ±2.0
| Benchmark | Result | Source | Date |
|---|---|---|---|
| Terminal-Bench 2.0 | 35.7% | Epoch AI | 2025-11-06 |
| METR task suite | 59.2% | Epoch AI | 2026-11-11 |
Reasoning 49.2 ±4.3
| Benchmark | Result | Source | Date |
|---|---|---|---|
| SimpleBench | 39.6% | SimpleBench | 2025-11-07 |
| GPQA Diamond (Vals run) | 78.5% | Vals AI | 2026-09-01 |
| Kagi Reasoning Benchmark | 64.4% | Kagi | 2026-11-11 |
Math 50.1 ±2.0
| Benchmark | Result | Source | Date |
|---|---|---|---|
| FrontierMath Tiers 1-3 (2025 set) | 21.4% | Epoch AI | 2025-12-05 |
| FrontierMath Tier 4 (2025 set) | 0.0% | Epoch AI | 2025-12-04 |
Knowledge 44.8 ±4.7
| Benchmark | Result | Source | Date |
|---|---|---|---|
| MMLU-Pro (Vals run) | 81.1% | Vals AI | 2026-09-01 |
General 56.1 ±3.0
| Benchmark | Result | Source | Date |
|---|---|---|---|
| EQ-Bench Creative Writing v3 | 1,631 | EQ-Bench | 2026-11-11 |
| EQ-Bench Longform Creative Writing | 69.0% | EQ-Bench | 2026-11-11 |
Names matched to this model
fireworks-models-kimi-k2-thinking, kimi-k2-thinking