Model comparison
Kimi K2 (Jul 2025) vs o3
o3 is the stronger model overall, scoring 47.5 to 41.2 on the Noometry Index. Kimi K2 (Jul 2025) costs 3.5× less per token, which makes it the better buy when o3's lead doesn't matter for your workload.
Last verified . 40 shared benchmarks.
Summary
- They share 40 benchmarks with published results for both. Kimi K2 (Jul 2025) scores higher in 0 categories and o3 in 9 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in knowledge, where o3 leads 54.6 to 37.3.
- The biggest single-benchmark swing is SimpleBench: 26.3% for Kimi K2 (Jul 2025) and 53.1% for o3.
- Kimi K2 (Jul 2025) is cheaper at $0.57 / $2.30 per million input/output tokens, against $2 / $8 for o3.
- Kimi K2 (Jul 2025) accepts more context: 262K tokens versus 200K.
- Kimi K2 (Jul 2025) has downloadable open weights; the other is API-only.
Side by side
| Kimi K2 (Jul 2025) | o3 | |
|---|---|---|
| Provider | Moonshot AI | OpenAI |
| Noometry Index | 41.2 | 47.5 |
| Released | 2025-07-12 | 2025-04-16 |
| Weights | Open | Proprietary |
| Context window | 262K | 200K |
| Max output | 262K | 100K |
| Input $ / M tokens | $0.57 | $2 |
| Output $ / M tokens | $2.30 | $8 |
| Results tracked | 42 | 63 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding o3 leads
Kimi K2 (Jul 2025): 42.4 (#102), o3: 46.8 (#64)
| Benchmark | Kimi K2 (Jul 2025) | o3 |
|---|---|---|
| SWE-bench Verified (bash only) | 63.4% | 58.4% |
| Aider Polyglot | 59.1% | 81.3% |
| GSO | 4.9% | 8.8% |
| WeirdML | 42.8% | 52.4% |
| LMArena Coding | 1399 | 1408 |
| ALE-Bench | 597.5 | 933.55 |
| SWE-bench Verified | — | 62.3% |
| CadEval | — | 74% |
Agentic & Tool Use o3 leads
Kimi K2 (Jul 2025): 32.4 (#64), o3: 34.5 (#44)
| Benchmark | Kimi K2 (Jul 2025) | o3 |
|---|---|---|
| Berkeley Function Calling Leaderboard | 59.1% | 63% |
| METR Time Horizons | 59.2% | 65.4% |
| Terminal-Bench | 35.7% | — |
| GDPval | — | 30.8% |
| DeepResearch Bench | — | 45.2% |
| OSWorld | — | 23% |
| LMArena Search | — | 1144 |
Reasoning o3 leads
Kimi K2 (Jul 2025): 23.3 (#179), o3: 32.0 (#78)
| Benchmark | Kimi K2 (Jul 2025) | o3 |
|---|---|---|
| SimpleBench | 26.3% | 53.1% |
| Kagi LLM Benchmark | 64.4% | 67.6% |
| LMArena Hard Prompts | 1384 | 1402 |
| Epoch Capabilities Index | 146.01 | 146.86 |
| ForecastBench | 60.2 | 62.5 |
| ARC-AGI-2 | — | 6.5% |
| ARC-AGI-1 | — | 60.8% |
| CritPt | — | 1.4% |
| Chess Puzzles | — | 38% |
| EnigmaEval | — | 13.1% |
| Mystery Game Puzzles | — | 29% |
| DTBench | — | 84.8% |
| LMCA | — | 39.7% |
Math o3 leads
Kimi K2 (Jul 2025): 42.7 (#83), o3: 50.2 (#58)
| Benchmark | Kimi K2 (Jul 2025) | o3 |
|---|---|---|
| Omni-MATH | 65.4% | 71.4% |
| LMArena Math | 1397 | 1426 |
| FrontierMath (Feb 2025 set) | 21.4% | 18.7% |
| FrontierMath Tier 4 (v1) | 0% | 2.1% |
| FrontierMath (Tiers 1-3) | — | 33.3% |
| OTIS Mock AIME 2024-2025 | — | 84.4% |
| MATH Level 5 | — | 97.8% |
Knowledge o3 leads
Kimi K2 (Jul 2025): 37.3 (#157), o3: 54.6 (#52)
| Benchmark | Kimi K2 (Jul 2025) | o3 |
|---|---|---|
| MMLU-Pro | 81.9% | 85.9% |
| Confabulations | 20.4% | 14.4% |
| GPQA (HELM) | 65.3% | 75.3% |
| LMArena Expert | 1365 | 1402 |
| GPQA Diamond | — | 81.8% |
| Humanity's Last Exam | — | 20.3% |
| SimpleQA Verified | — | 49.4% |
| Vectara Hallucination Rate | 17.9% | — |
Multimodal Not comparable
Kimi K2 (Jul 2025): —, o3: 41.4 (#36)
| Benchmark | Kimi K2 (Jul 2025) | o3 |
|---|---|---|
| LMArena Vision | — | 1214 |
| GeoBench | — | 74% |
| VPCT | — | 52% |
Multilingual o3 leads
Kimi K2 (Jul 2025): 49.6 (#130), o3: 51.7 (#105)
| Benchmark | Kimi K2 (Jul 2025) | o3 |
|---|---|---|
| LMArena Non-English | 1372 | 1401 |
| LMArena Chinese | 1415 | 1437 |
| LMArena French | 1379 | 1430 |
| LMArena German | 1387 | 1420 |
| LMArena Japanese | 1349 | 1403 |
| LMArena Korean | 1325 | 1370 |
| LMArena Russian | 1385 | 1406 |
| LMArena Spanish | 1386 | 1395 |
Instruction Following o3 leads
Kimi K2 (Jul 2025): 71.1 (#156), o3: 72.8 (#127)
| Benchmark | Kimi K2 (Jul 2025) | o3 |
|---|---|---|
| IFEval | 85% | 86.9% |
| LMArena Instruction Following | 1348 | 1368 |
Long Context o3 leads
Kimi K2 (Jul 2025): 41.2 (#145), o3: 53.3 (#6)
| Benchmark | Kimi K2 (Jul 2025) | o3 |
|---|---|---|
| Fiction.LiveBench | 66.7% | 88.9% |
| CL-bench | 17.6% | 17.8% |
| LMArena Longer Query | 1353 | 1372 |
Writing & Preference o3 leads
Kimi K2 (Jul 2025): 62.3 (#78), o3: 63.5 (#64)
| Benchmark | Kimi K2 (Jul 2025) | o3 |
|---|---|---|
| LMArena Text | 1380 | 1410 |
| LMArena Creative Writing | 1350 | 1359 |
| Short-Story Creative Writing | 85.6% | 83.9% |
| EQ-Bench Creative Writing | 1666 | 1676 |
| WildBench | 86.2% | 86.1% |
| LMArena Multi-Turn | 1371 | 1405 |
Frequently asked questions
Is Kimi K2 (Jul 2025) better than o3?
o3 is the stronger model overall, scoring 47.5 to 41.2 on the Noometry Index. Kimi K2 (Jul 2025) costs 3.5× less per token, which makes it the better buy when o3's lead doesn't matter for your workload.
Which is cheaper, Kimi K2 (Jul 2025) or o3?
Kimi K2 (Jul 2025) is cheaper. It lists at $0.57 per million input tokens and $2.30 per million output tokens; o3 lists at $2 and $8.
Is Kimi K2 (Jul 2025) or o3 better for coding?
o3 scores higher on coding benchmarks: 46.8 versus 42.4 in the Noometry coding category.
Which has the bigger context window?
Kimi K2 (Jul 2025) does, with 262K tokens against 200K.
How many benchmarks do Kimi K2 (Jul 2025) and o3 share?
40 benchmarks have published results for both models. Kimi K2 (Jul 2025) has 42 scored results on Noometry and o3 has 63.