Model comparison
Grok 3 vs Kimi K2 (Jul 2025)
Kimi K2 (Jul 2025) is the stronger model overall, scoring 41.2 to 39.9 on the Noometry Index.
Last verified . 34 shared benchmarks.
Summary
- They share 34 benchmarks with published results for both. Grok 3 scores higher in 3 categories and Kimi K2 (Jul 2025) in 6 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where Kimi K2 (Jul 2025) leads 23.3 to 13.7.
- The biggest single-benchmark swing is Omni-MATH: 46.4% for Grok 3 and 65.4% for Kimi K2 (Jul 2025).
- Kimi K2 (Jul 2025) has downloadable open weights; the other is API-only.
Side by side
| Grok 3 | Kimi K2 (Jul 2025) | |
|---|---|---|
| Provider | xAI | Moonshot AI |
| Noometry Index | 39.9 | 41.2 |
| Released | 2025-04-09 | 2025-07-12 |
| Weights | Proprietary | Open |
| Context window | — | 262K |
| Max output | — | 262K |
| Input $ / M tokens | — | $0.57 |
| Output $ / M tokens | — | $2.30 |
| Results tracked | 40 | 42 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Too close to call
Grok 3: 41.9 (#115), Kimi K2 (Jul 2025): 42.4 (#102)
| Benchmark | Grok 3 | Kimi K2 (Jul 2025) |
|---|---|---|
| Aider Polyglot | 53.3% | 59.1% |
| WeirdML | 37.2% | 42.8% |
| LMArena Coding | 1432 | 1399 |
| SWE-bench Verified (bash only) | — | 63.4% |
| GSO | — | 4.9% |
| ALE-Bench | — | 597.5 |
Agentic & Tool Use Kimi K2 (Jul 2025) leads
Grok 3: 30.5 (#76), Kimi K2 (Jul 2025): 32.4 (#64)
| Benchmark | Grok 3 | Kimi K2 (Jul 2025) |
|---|---|---|
| Terminal-Bench | — | 35.7% |
| Berkeley Function Calling Leaderboard | — | 59.1% |
| BALROG | 29.5% | — |
| METR Time Horizons | — | 59.2% |
Reasoning Kimi K2 (Jul 2025) leads
Grok 3: 13.7 (#333), Kimi K2 (Jul 2025): 23.3 (#179)
| Benchmark | Grok 3 | Kimi K2 (Jul 2025) |
|---|---|---|
| SimpleBench | 36.1% | 26.3% |
| Kagi LLM Benchmark | 61.3% | 64.4% |
| LMArena Hard Prompts | 1434 | 1384 |
| Epoch Capabilities Index | 138.33 | 146.01 |
| ARC-AGI-2 | 0% | — |
| ARC-AGI-1 | 5.5% | — |
| ForecastBench | — | 60.2 |
Math Kimi K2 (Jul 2025) leads
Grok 3: 38.0 (#145), Kimi K2 (Jul 2025): 42.7 (#83)
| Benchmark | Grok 3 | Kimi K2 (Jul 2025) |
|---|---|---|
| Omni-MATH | 46.4% | 65.4% |
| LMArena Math | 1391 | 1397 |
| FrontierMath (Feb 2025 set) | 3.8% | 21.4% |
| FrontierMath Tier 4 (v1) | 0% | 0% |
| OTIS Mock AIME 2024-2025 | 55.6% | — |
| MATH Level 5 | 88.7% | — |
Knowledge Grok 3 leads
Grok 3: 46.2 (#82), Kimi K2 (Jul 2025): 37.3 (#157)
| Benchmark | Grok 3 | Kimi K2 (Jul 2025) |
|---|---|---|
| MMLU-Pro | 78.8% | 81.9% |
| Confabulations | 14.2% | 20.4% |
| Vectara Hallucination Rate | 5.8% | 17.9% |
| GPQA (HELM) | 65% | 65.3% |
| LMArena Expert | 1421 | 1365 |
| GPQA Diamond | 75.8% | — |
Multilingual Grok 3 leads
Grok 3: 52.3 (#87), Kimi K2 (Jul 2025): 49.6 (#130)
| Benchmark | Grok 3 | Kimi K2 (Jul 2025) |
|---|---|---|
| LMArena Non-English | 1410 | 1372 |
| LMArena Chinese | 1448 | 1415 |
| LMArena French | 1460 | 1379 |
| LMArena German | 1431 | 1387 |
| LMArena Japanese | 1387 | 1349 |
| LMArena Korean | 1373 | 1325 |
| LMArena Russian | 1416 | 1385 |
| LMArena Spanish | 1417 | 1386 |
Instruction Following Grok 3 leads
Grok 3: 75.0 (#73), Kimi K2 (Jul 2025): 71.1 (#156)
| Benchmark | Grok 3 | Kimi K2 (Jul 2025) |
|---|---|---|
| IFEval | 88.4% | 85% |
| LMArena Instruction Following | 1409 | 1348 |
Long Context Kimi K2 (Jul 2025) leads
Grok 3: 38.7 (#192), Kimi K2 (Jul 2025): 41.2 (#145)
| Benchmark | Grok 3 | Kimi K2 (Jul 2025) |
|---|---|---|
| Fiction.LiveBench | 58.3% | 66.7% |
| LMArena Longer Query | 1439 | 1353 |
| CL-bench | — | 17.6% |
Writing & Preference Kimi K2 (Jul 2025) leads
Grok 3: 55.8 (#141), Kimi K2 (Jul 2025): 62.3 (#78)
| Benchmark | Grok 3 | Kimi K2 (Jul 2025) |
|---|---|---|
| LMArena Text | 1426 | 1380 |
| LMArena Creative Writing | 1414 | 1350 |
| Short-Story Creative Writing | 76.4% | 85.6% |
| EQ-Bench Creative Writing | 1186 | 1666 |
| WildBench | 84.9% | 86.2% |
| LMArena Multi-Turn | 1425 | 1371 |
Frequently asked questions
Is Grok 3 better than Kimi K2 (Jul 2025)?
Kimi K2 (Jul 2025) is the stronger model overall, scoring 41.2 to 39.9 on the Noometry Index.
Is Grok 3 or Kimi K2 (Jul 2025) better for coding?
They score almost the same on coding (41.9 vs 42.4); test both on your own repository before choosing.
How many benchmarks do Grok 3 and Kimi K2 (Jul 2025) share?
34 benchmarks have published results for both models. Grok 3 has 40 scored results on Noometry and Kimi K2 (Jul 2025) has 42.