Model comparison
Claude 3.5 Haiku vs Kimi K2 (Jul 2025)
Kimi K2 (Jul 2025) is the stronger model overall, scoring 40.9 to 28.9 on the Noometry Index.
Last verified . 29 shared benchmarks.
Summary
- They share 29 benchmarks with published results for both. Claude 3.5 Haiku scores higher in 0 categories and Kimi K2 (Jul 2025) in 9 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in math, where Kimi K2 (Jul 2025) leads 40.1 to 13.7.
- The biggest single-benchmark swing is Omni-MATH: 22.4% for Claude 3.5 Haiku and 65.4% for Kimi K2 (Jul 2025).
- Kimi K2 (Jul 2025) has downloadable open weights; the other is API-only.
Side by side
| Claude 3.5 Haiku | Kimi K2 (Jul 2025) | |
|---|---|---|
| Provider | Anthropic | Moonshot AI |
| Noometry Index | 28.9 | 40.9 |
| Released | 2024-10-22 | 2025-07-12 |
| Weights | Proprietary | Open |
| Context window | — | 262K |
| Max output | — | 262K |
| Input $ / M tokens | — | $0.57 |
| Output $ / M tokens | — | $2.30 |
| Results tracked | 49 | 42 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Kimi K2 (Jul 2025) leads
Claude 3.5 Haiku: 32.6 (#258), Kimi K2 (Jul 2025): 43.1 (#83)
| Benchmark | Claude 3.5 Haiku | Kimi K2 (Jul 2025) |
|---|---|---|
| Aider Polyglot | 28% | 59.1% |
| WeirdML | 30.7% | 42.8% |
| LMArena Coding | 1287 | 1399 |
| SWE-bench Verified (bash only) | — | 63.4% |
| SciCode | 27.4% | — |
| GSO | — | 4.9% |
| BigCodeBench Instruct | 46.1% | — |
| LiveBench Coding | 51.4% | — |
| BigCodeBench Complete | 59% | — |
| CadEval | 32% | — |
| ALE-Bench | — | 597.5 |
Agentic & Tool Use Kimi K2 (Jul 2025) leads
Claude 3.5 Haiku: 27.7 (#90), Kimi K2 (Jul 2025): 32.2 (#59)
| Benchmark | Claude 3.5 Haiku | Kimi K2 (Jul 2025) |
|---|---|---|
| Terminal-Bench | — | 35.7% |
| Berkeley Function Calling Leaderboard | — | 59.1% |
| BALROG | 19.3% | — |
| METR Time Horizons | — | 59.2% |
Reasoning Kimi K2 (Jul 2025) leads
Claude 3.5 Haiku: 17.1 (#286), Kimi K2 (Jul 2025): 22.7 (#173)
| Benchmark | Claude 3.5 Haiku | Kimi K2 (Jul 2025) |
|---|---|---|
| LMArena Hard Prompts | 1251 | 1384 |
| Epoch Capabilities Index | 127.15 | 146.02 |
| SimpleBench | — | 26.3% |
| Kagi LLM Benchmark | — | 64.4% |
| CritPt | 0% | — |
| LiveBench Reasoning | 28.1% | — |
| DTBench | 56.7% | — |
| LiveBench Data Analysis | 48.5% | — |
| ForecastBench | — | 60.2 |
| LiveBench | 43.5% | — |
Math Kimi K2 (Jul 2025) leads
Claude 3.5 Haiku: 13.7 (#294), Kimi K2 (Jul 2025): 40.1 (#81)
| Benchmark | Claude 3.5 Haiku | Kimi K2 (Jul 2025) |
|---|---|---|
| Omni-MATH | 22.4% | 65.4% |
| LMArena Math | 1244 | 1396 |
| FrontierMath (Feb 2025 set) | 0.3% | 21.4% |
| OTIS Mock AIME 2024-2025 | 4.3% | — |
| LiveBench Math | 35.5% | — |
| MATH Level 5 | 46.4% | — |
| FrontierMath Tier 4 (v1) | — | 0% |
Knowledge Kimi K2 (Jul 2025) leads
Claude 3.5 Haiku: 18.7 (#276), Kimi K2 (Jul 2025): 36.9 (#153)
| Benchmark | Claude 3.5 Haiku | Kimi K2 (Jul 2025) |
|---|---|---|
| MMLU-Pro | 60.5% | 81.9% |
| Confabulations | 36.7% | 20.4% |
| GPQA (HELM) | 36.3% | 65.3% |
| LMArena Expert | 1208 | 1363 |
| GPQA Diamond | 38.1% | — |
| Vectara Hallucination Rate | — | 17.9% |
| MMLU | 74.3% | — |
Multimodal Not comparable
Claude 3.5 Haiku: 26.3 (#113), Kimi K2 (Jul 2025): —
| Benchmark | Claude 3.5 Haiku | Kimi K2 (Jul 2025) |
|---|---|---|
| LMArena Vision | 1092 | — |
| GeoBench | 34% | — |
Multilingual Kimi K2 (Jul 2025) leads
Claude 3.5 Haiku: 40.2 (#213), Kimi K2 (Jul 2025): 49.8 (#125)
| Benchmark | Claude 3.5 Haiku | Kimi K2 (Jul 2025) |
|---|---|---|
| LMArena Non-English | 1238 | 1372 |
| LMArena Chinese | 1229 | 1416 |
| LMArena French | 1262 | 1379 |
| LMArena German | 1237 | 1386 |
| LMArena Japanese | 1175 | 1349 |
| LMArena Korean | 1172 | 1324 |
| LMArena Russian | 1253 | 1384 |
| LMArena Spanish | 1261 | 1383 |
Instruction Following Kimi K2 (Jul 2025) leads
Claude 3.5 Haiku: 63.1 (#229), Kimi K2 (Jul 2025): 71.3 (#151)
| Benchmark | Claude 3.5 Haiku | Kimi K2 (Jul 2025) |
|---|---|---|
| IFEval | 79.2% | 85% |
| LMArena Instruction Following | 1241 | 1347 |
| LiveBench Instruction Following | 61.9% | — |
Long Context Kimi K2 (Jul 2025) leads
Claude 3.5 Haiku: 38.3 (#195), Kimi K2 (Jul 2025): 41.2 (#140)
| Benchmark | Claude 3.5 Haiku | Kimi K2 (Jul 2025) |
|---|---|---|
| LMArena Longer Query | 1261 | 1353 |
| Fiction.LiveBench | — | 66.7% |
| CL-bench | — | 17.6% |
Writing & Preference Kimi K2 (Jul 2025) leads
Claude 3.5 Haiku: 42.9 (#228), Kimi K2 (Jul 2025): 62.5 (#76)
| Benchmark | Claude 3.5 Haiku | Kimi K2 (Jul 2025) |
|---|---|---|
| LMArena Text | 1255 | 1380 |
| LMArena Creative Writing | 1233 | 1349 |
| Short-Story Creative Writing | 73.5% | 85.6% |
| EQ-Bench Creative Writing | 1146 | 1666 |
| WildBench | 76% | 86.2% |
| LMArena Multi-Turn | 1265 | 1371 |
| LiveBench Language | 35.4% | — |
Frequently asked questions
Is Claude 3.5 Haiku better than Kimi K2 (Jul 2025)?
Kimi K2 (Jul 2025) is the stronger model overall, scoring 40.9 to 28.9 on the Noometry Index.
Is Claude 3.5 Haiku or Kimi K2 (Jul 2025) better for coding?
Kimi K2 (Jul 2025) scores higher on coding benchmarks: 43.1 versus 32.6 in the Noometry coding category.
How many benchmarks do Claude 3.5 Haiku and Kimi K2 (Jul 2025) share?
29 benchmarks have published results for both models. Claude 3.5 Haiku has 49 scored results on Noometry and Kimi K2 (Jul 2025) has 42.