Model comparison
DeepSeek-V3.1 vs Kimi K2 Thinking Turbo
Kimi K2 Thinking Turbo is the stronger model overall, scoring 45.8 to 42.8 on the Noometry Index.
Last verified . 17 shared benchmarks.
DeepSeekSummary
- They share 17 benchmarks with published results for both. DeepSeek-V3.1 scores higher in 3 categories and Kimi K2 Thinking Turbo in 5 categories; 5 gaps are clear of the uncertainty.
- The widest gap is in math, where Kimi K2 Thinking Turbo leads 47.4 to 38.9.
Side by side
| DeepSeek-V3.1 | Kimi K2 Thinking Turbo | |
|---|---|---|
| Provider | DeepSeek | Moonshot AI |
| Noometry Index | 42.8 | 45.8 |
| Released | 2025-08-21 | 2025-11-06 |
| Weights | Open | Open |
| Context window | 164K | — |
| Max output | 8K | — |
| Input $ / M tokens | $0.25 | — |
| Output $ / M tokens | $0.95 | — |
| Results tracked | 27 | 21 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding DeepSeek-V3.1 leads
DeepSeek-V3.1: 40.3 (#144), Kimi K2 Thinking Turbo: 38.4 (#178)
| Benchmark | DeepSeek-V3.1 | Kimi K2 Thinking Turbo |
|---|---|---|
| LMArena Coding | 1417 | 1454 |
| LMArena WebDev | — | 1322 |
| WeirdML | 38.4% | — |
Reasoning Kimi K2 Thinking Turbo leads
DeepSeek-V3.1: 27.9 (#110), Kimi K2 Thinking Turbo: 32.5 (#75)
| Benchmark | DeepSeek-V3.1 | Kimi K2 Thinking Turbo |
|---|---|---|
| LMArena Hard Prompts | 1417 | 1428 |
| SimpleBench | 40% | — |
| Kagi LLM Benchmark | 53.2% | — |
| Chess Puzzles | — | 20% |
| DTBench | 82.7% | — |
| LMCA | 24.3% | — |
| Epoch Capabilities Index | 139.92 | — |
| ForecastBench | 58 | — |
Math Kimi K2 Thinking Turbo leads
DeepSeek-V3.1: 38.9 (#122), Kimi K2 Thinking Turbo: 47.4 (#68)
| Benchmark | DeepSeek-V3.1 | Kimi K2 Thinking Turbo |
|---|---|---|
| LMArena Math | 1420 | 1429 |
| OTIS Mock AIME 2024-2025 | — | 83.1% |
Knowledge Kimi K2 Thinking Turbo leads
DeepSeek-V3.1: 43.7 (#90), Kimi K2 Thinking Turbo: 50.9 (#69)
| Benchmark | DeepSeek-V3.1 | Kimi K2 Thinking Turbo |
|---|---|---|
| LMArena Expert | 1405 | 1439 |
| GPQA Diamond | — | 84.2% |
| Vectara Hallucination Rate | 5.5% | — |
Multilingual Too close to call
DeepSeek-V3.1: 51.6 (#106), Kimi K2 Thinking Turbo: 51.4 (#109)
| Benchmark | DeepSeek-V3.1 | Kimi K2 Thinking Turbo |
|---|---|---|
| LMArena Non-English | 1400 | 1398 |
| LMArena Chinese | 1469 | 1456 |
| LMArena French | 1447 | 1425 |
| LMArena German | 1411 | 1390 |
| LMArena Japanese | 1378 | 1357 |
| LMArena Korean | 1337 | 1330 |
| LMArena Russian | 1405 | 1391 |
| LMArena Spanish | 1431 | 1406 |
Instruction Following Too close to call
DeepSeek-V3.1: 73.9 (#110), Kimi K2 Thinking Turbo: 74.0 (#109)
| Benchmark | DeepSeek-V3.1 | Kimi K2 Thinking Turbo |
|---|---|---|
| LMArena Instruction Following | 1400 | 1403 |
Long Context Kimi K2 Thinking Turbo leads
DeepSeek-V3.1: 36.3 (#232), Kimi K2 Thinking Turbo: 43.2 (#102)
| Benchmark | DeepSeek-V3.1 | Kimi K2 Thinking Turbo |
|---|---|---|
| LMArena Longer Query | 1422 | 1415 |
| Fiction.LiveBench | 52.8% | — |
Writing & Preference Too close to call
DeepSeek-V3.1: 60.3 (#98), Kimi K2 Thinking Turbo: 60.0 (#104)
| Benchmark | DeepSeek-V3.1 | Kimi K2 Thinking Turbo |
|---|---|---|
| LMArena Text | 1420 | 1415 |
| LMArena Creative Writing | 1401 | 1374 |
| LMArena Multi-Turn | 1408 | 1414 |
| EQ-Bench Creative Writing | 1436 | — |
Frequently asked questions
Is DeepSeek-V3.1 better than Kimi K2 Thinking Turbo?
Kimi K2 Thinking Turbo is the stronger model overall, scoring 45.8 to 42.8 on the Noometry Index.
Is DeepSeek-V3.1 or Kimi K2 Thinking Turbo better for coding?
DeepSeek-V3.1 scores higher on coding benchmarks: 40.3 versus 38.4 in the Noometry coding category.
How many benchmarks do DeepSeek-V3.1 and Kimi K2 Thinking Turbo share?
17 benchmarks have published results for both models. DeepSeek-V3.1 has 27 scored results on Noometry and Kimi K2 Thinking Turbo has 21.