Model comparison
Granite 4.0 Micro vs Kimi K2 Thinking Turbo
Kimi K2 Thinking Turbo is the stronger model overall, scoring 45.8 to 29.0 on the Noometry Index.
Last verified . 3 shared benchmarks.
Summary
- They share 3 benchmarks with published results for both. Granite 4.0 Micro scores higher in 0 categories and Kimi K2 Thinking Turbo in 5 categories; 5 gaps are clear of the uncertainty.
- The widest gap is in knowledge, where Kimi K2 Thinking Turbo leads 50.9 to 9.9.
- The biggest single-benchmark swing is OTIS Mock AIME 2024-2025: 2.8% for Granite 4.0 Micro and 83.1% for Kimi K2 Thinking Turbo.
Side by side
| Granite 4.0 Micro | Kimi K2 Thinking Turbo | |
|---|---|---|
| Provider | IBM | Moonshot AI |
| Noometry Index | 29.0 | 45.8 |
| Released | 2025-10-02 | 2025-11-06 |
| Weights | Open | Open |
| Context window | 131K | — |
| Max output | 118K | — |
| Input $ / M tokens | $0.017 | — |
| Output $ / M tokens | $0.11 | — |
| Results tracked | 8 | 21 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Not comparable
Granite 4.0 Micro: —, Kimi K2 Thinking Turbo: 38.4 (#178)
| Benchmark | Granite 4.0 Micro | Kimi K2 Thinking Turbo |
|---|---|---|
| LMArena WebDev | — | 1322 |
| LMArena Coding | — | 1454 |
Reasoning Kimi K2 Thinking Turbo leads
Granite 4.0 Micro: 19.2 (#265), Kimi K2 Thinking Turbo: 32.5 (#75)
| Benchmark | Granite 4.0 Micro | Kimi K2 Thinking Turbo |
|---|---|---|
| Chess Puzzles | 0% | 20% |
| LMArena Hard Prompts | — | 1428 |
Math Kimi K2 Thinking Turbo leads
Granite 4.0 Micro: 12.0 (#307), Kimi K2 Thinking Turbo: 47.4 (#68)
| Benchmark | Granite 4.0 Micro | Kimi K2 Thinking Turbo |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 2.8% | 83.1% |
| Omni-MATH | 20.9% | — |
| LMArena Math | — | 1429 |
Knowledge Kimi K2 Thinking Turbo leads
Granite 4.0 Micro: 9.9 (#304), Kimi K2 Thinking Turbo: 50.9 (#69)
| Benchmark | Granite 4.0 Micro | Kimi K2 Thinking Turbo |
|---|---|---|
| GPQA Diamond | 28.3% | 84.2% |
| MMLU-Pro | 39.5% | — |
| GPQA (HELM) | 30.7% | — |
| LMArena Expert | — | 1439 |
Multilingual Not comparable
Granite 4.0 Micro: —, Kimi K2 Thinking Turbo: 51.4 (#109)
| Benchmark | Granite 4.0 Micro | Kimi K2 Thinking Turbo |
|---|---|---|
| LMArena Non-English | — | 1398 |
| LMArena Chinese | — | 1456 |
| LMArena French | — | 1425 |
| LMArena German | — | 1390 |
| LMArena Japanese | — | 1357 |
| LMArena Korean | — | 1330 |
| LMArena Russian | — | 1391 |
| LMArena Spanish | — | 1406 |
Instruction Following Kimi K2 Thinking Turbo leads
Granite 4.0 Micro: 69.9 (#169), Kimi K2 Thinking Turbo: 74.0 (#109)
| Benchmark | Granite 4.0 Micro | Kimi K2 Thinking Turbo |
|---|---|---|
| IFEval | 84.9% | — |
| LMArena Instruction Following | — | 1403 |
Long Context Not comparable
Granite 4.0 Micro: —, Kimi K2 Thinking Turbo: 43.2 (#102)
| Benchmark | Granite 4.0 Micro | Kimi K2 Thinking Turbo |
|---|---|---|
| LMArena Longer Query | — | 1415 |
Writing & Preference Kimi K2 Thinking Turbo leads
Granite 4.0 Micro: 46.7 (#216), Kimi K2 Thinking Turbo: 60.0 (#104)
| Benchmark | Granite 4.0 Micro | Kimi K2 Thinking Turbo |
|---|---|---|
| LMArena Text | — | 1415 |
| LMArena Creative Writing | — | 1374 |
| WildBench | 67% | — |
| LMArena Multi-Turn | — | 1414 |
Frequently asked questions
Is Granite 4.0 Micro better than Kimi K2 Thinking Turbo?
Kimi K2 Thinking Turbo is the stronger model overall, scoring 45.8 to 29.0 on the Noometry Index.
How many benchmarks do Granite 4.0 Micro and Kimi K2 Thinking Turbo share?
3 benchmarks have published results for both models. Granite 4.0 Micro has 8 scored results on Noometry and Kimi K2 Thinking Turbo has 21.