Model comparison
Claude Sonnet 4 vs Kimi K2.6
Kimi K2.6 is the stronger model overall, scoring 47.5 to 40.6 on the Noometry Index.
Last verified . 31 shared benchmarks.
Summary
- They share 31 benchmarks with published results for both. Claude Sonnet 4 scores higher in 1 category and Kimi K2.6 in 9 categories; 10 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where Kimi K2.6 leads 39.6 to 23.0.
- The biggest single-benchmark swing is OTIS Mock AIME 2024-2025: 71.1% for Claude Sonnet 4 and 96.1% for Kimi K2.6.
- Kimi K2.6 is cheaper at $0.95 / $4 per million input/output tokens, against $3 / $15 for Claude Sonnet 4.
- Kimi K2.6 accepts more context: 262K tokens versus 200K.
- Kimi K2.6 has downloadable open weights; the other is API-only.
Side by side
| Claude Sonnet 4 | Kimi K2.6 | |
|---|---|---|
| Provider | Anthropic | Moonshot AI |
| Noometry Index | 40.6 | 47.5 |
| Released | 2025-05-22 | 2026-04-20 |
| Weights | Proprietary | Open |
| Context window | 200K | 262K |
| Max output | 64K | 262K |
| Input $ / M tokens | $3 | $0.95 |
| Output $ / M tokens | $15 | $4 |
| Results tracked | 58 | 51 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Kimi K2.6 leads
Claude Sonnet 4: 44.0 (#70), Kimi K2.6: 50.5 (#41)
| Benchmark | Claude Sonnet 4 | Kimi K2.6 |
|---|---|---|
| SciCode | 40% | 53.5% |
| WeirdML | 46.1% | 55.9% |
| LMArena Coding | 1414 | 1488 |
| ALE-Bench | 655.35 | 1,093 |
| SWE-bench Verified | — | 76.7% |
| SWE-bench Verified (bash only) | 64.9% | — |
| Aider Polyglot | 61.3% | — |
| LMArena WebDev | — | 1509 |
| GSO | 4.9% | — |
Agentic & Tool Use Claude Sonnet 4 leads
Claude Sonnet 4: 38.3 (#29), Kimi K2.6: 21.8 (#128)
| Benchmark | Claude Sonnet 4 | Kimi K2.6 |
|---|---|---|
| OSWorld 2.0 | — | 4.6% |
| TheAgentCompany | 33.1% | — |
| Cybench | 35% | — |
| DeepResearch Bench | 46.6% | — |
| OSWorld | 43.9% | — |
| ExploitBench | — | 18.4% |
| GBAEval | — | 0.9% |
| GDP.pdf | — | 12% |
| METR Time Horizons | 62% | — |
| Vending-Bench 2 | — | 6,205 |
Reasoning Kimi K2.6 leads
Claude Sonnet 4: 23.0 (#169), Kimi K2.6: 39.6 (#51)
| Benchmark | Claude Sonnet 4 | Kimi K2.6 |
|---|---|---|
| CritPt | 0.3% | 8% |
| LMArena Hard Prompts | 1372 | 1470 |
| DTBench | 77.1% | 90.9% |
| LMCA | 29% | 37.3% |
| Epoch Capabilities Index | 141.69 | 151.03 |
| ARC-AGI-2 | 5.9% | — |
| SimpleBench | 45.5% | — |
| Kagi LLM Benchmark | 73% | — |
| NYT Connections (extended) | — | 87.2% |
| ARC-AGI-1 | 40% | — |
| Chess Puzzles | — | 26% |
| EnigmaEval | 3.1% | — |
| EBR-Bench | — | 2.4% |
| Mystery Game Puzzles | — | 18% |
| ForecastBench | 60.2 | — |
Math Kimi K2.6 leads
Claude Sonnet 4: 41.5 (#71), Kimi K2.6: 56.8 (#34)
| Benchmark | Claude Sonnet 4 | Kimi K2.6 |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 71.1% | 96.1% |
| LMArena Math | 1375 | 1475 |
| FrontierMath (Feb 2025 set) | 4.1% | 39% |
| FrontierMath Tier 4 (v1) | 0% | 14.6% |
| FrontierMath (Tiers 1-3) | — | 57.2% |
| FrontierMath Tier 4 | — | 25.6% |
| MathArena Final-Answer Competitions | — | 72.9% |
| ProofBench | — | 16% |
| Omni-MATH | 60.2% | — |
| MATH Level 5 | 84.4% | — |
Knowledge Kimi K2.6 leads
Claude Sonnet 4: 41.6 (#99), Kimi K2.6: 53.9 (#48)
| Benchmark | Claude Sonnet 4 | Kimi K2.6 |
|---|---|---|
| GPQA Diamond | 79.2% | 90.8% |
| Vectara Hallucination Rate | 10.3% | 10.8% |
| LMArena Expert | 1371 | 1491 |
| Humanity's Last Exam | 7.8% | — |
| SimpleQA Verified | — | 34.9% |
| MMLU-Pro | 84.3% | — |
| Confabulations | 13.2% | — |
| GPQA (HELM) | 70.6% | — |
Multimodal Kimi K2.6 leads
Claude Sonnet 4: 25.6 (#116), Kimi K2.6: 31.8 (#95)
| Benchmark | Claude Sonnet 4 | Kimi K2.6 |
|---|---|---|
| LMArena Vision | 1191 | 1282 |
| GeoBench | 37% | — |
| VPCT | 34% | — |
| Blueprint-Bench 2 | — | 3.9% |
| Furniture Assembly | — | 21.7% |
| LMArena Document | — | 1451 |
| MindCube | 44.8% | — |
Multilingual Kimi K2.6 leads
Claude Sonnet 4: 46.9 (#151), Kimi K2.6: 55.1 (#35)
| Benchmark | Claude Sonnet 4 | Kimi K2.6 |
|---|---|---|
| LMArena Non-English | 1333 | 1446 |
| LMArena Chinese | 1351 | 1521 |
| LMArena French | 1364 | 1471 |
| LMArena German | 1331 | 1449 |
| LMArena Japanese | 1301 | 1442 |
| LMArena Korean | 1293 | 1428 |
| LMArena Russian | 1355 | 1446 |
| LMArena Spanish | 1356 | 1463 |
Instruction Following Kimi K2.6 leads
Claude Sonnet 4: 71.8 (#141), Kimi K2.6: 76.5 (#40)
| Benchmark | Claude Sonnet 4 | Kimi K2.6 |
|---|---|---|
| LMArena Instruction Following | 1375 | 1451 |
| IFEval | 84% | — |
Long Context Kimi K2.6 leads
Claude Sonnet 4: 33.6 (#254), Kimi K2.6: 44.9 (#50)
| Benchmark | Claude Sonnet 4 | Kimi K2.6 |
|---|---|---|
| LMArena Longer Query | 1398 | 1468 |
| Fiction.LiveBench | 46.9% | — |
Writing & Preference Kimi K2.6 leads
Claude Sonnet 4: 57.3 (#127), Kimi K2.6: 68.7 (#25)
| Benchmark | Claude Sonnet 4 | Kimi K2.6 |
|---|---|---|
| LMArena Text | 1351 | 1455 |
| LMArena Creative Writing | 1345 | 1434 |
| EQ-Bench Creative Writing | 1483 | 1725 |
| LMArena Multi-Turn | 1376 | 1453 |
| Short-Story Creative Writing | 81.4% | — |
| WildBench | 83.8% | — |
| EQ-Bench 4 | — | 1202 |
Frequently asked questions
Is Claude Sonnet 4 better than Kimi K2.6?
Kimi K2.6 is the stronger model overall, scoring 47.5 to 40.6 on the Noometry Index.
Which is cheaper, Claude Sonnet 4 or Kimi K2.6?
Kimi K2.6 is cheaper. It lists at $0.95 per million input tokens and $4 per million output tokens; Claude Sonnet 4 lists at $3 and $15.
Is Claude Sonnet 4 or Kimi K2.6 better for coding?
Kimi K2.6 scores higher on coding benchmarks: 50.5 versus 44.0 in the Noometry coding category.
Which has the bigger context window?
Kimi K2.6 does, with 262K tokens against 200K.
How many benchmarks do Claude Sonnet 4 and Kimi K2.6 share?
31 benchmarks have published results for both models. Claude Sonnet 4 has 58 scored results on Noometry and Kimi K2.6 has 51.