Model comparison
Claude Opus 4 vs Kimi K2.5
Kimi K2.5 is the stronger model overall, scoring 48.1 to 43.1 on the Noometry Index.
Last verified . 36 shared benchmarks.
Summary
- They share 36 benchmarks with published results for both. Claude Opus 4 scores higher in 2 categories and Kimi K2.5 in 8 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in long context, where Kimi K2.5 leads 52.1 to 39.6.
- The biggest single-benchmark swing is ARC-AGI-1: 35.7% for Claude Opus 4 and 65.3% for Kimi K2.5.
- Kimi K2.5 is cheaper at $0.45 / $2.25 per million input/output tokens, against $15 / $75 for Claude Opus 4.
- Kimi K2.5 accepts more context: 262K tokens versus 200K.
- Kimi K2.5 has downloadable open weights; the other is API-only.
Side by side
| Claude Opus 4 | Kimi K2.5 | |
|---|---|---|
| Provider | Anthropic | Moonshot AI |
| Noometry Index | 43.1 | 48.1 |
| Released | 2025-05-22 | 2026-01-27 |
| Weights | Proprietary | Open |
| Context window | 200K | 262K |
| Max output | 32K | 262K |
| Input $ / M tokens | $15 | $0.45 |
| Output $ / M tokens | $75 | $2.25 |
| Results tracked | 56 | 51 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Kimi K2.5 leads
Claude Opus 4: 47.2 (#62), Kimi K2.5: 48.8 (#53)
| Benchmark | Claude Opus 4 | Kimi K2.5 |
|---|---|---|
| SWE-bench Verified | 70.7% | 73.8% |
| SWE-bench Verified (bash only) | 67.6% | 70.8% |
| WeirdML | 43.7% | 45.6% |
| LMArena Coding | 1442 | 1474 |
| Aider Polyglot | 72% | — |
| LMArena WebDev | — | 1437 |
| SWE-bench Multilingual | — | 67.3% |
| SciCode | — | 49% |
| GSO | 6.9% | — |
| ALE-Bench | — | 821.65 |
| AlgoTune | 1.33 | — |
Agentic & Tool Use Too close to call
Claude Opus 4: 34.8 (#42), Kimi K2.5: 34.2 (#48)
| Benchmark | Claude Opus 4 | Kimi K2.5 |
|---|---|---|
| Terminal-Bench | — | 43.2% |
| Cybench | 38% | — |
| DeepResearch Bench | 46.8% | — |
| OSWorld | — | 63.3% |
| LMArena Search | 1127 | — |
| METR Time Horizons | 63.9% | — |
| Vending-Bench 2 | — | 1,198 |
Reasoning Kimi K2.5 leads
Claude Opus 4: 27.3 (#121), Kimi K2.5: 31.2 (#80)
| Benchmark | Claude Opus 4 | Kimi K2.5 |
|---|---|---|
| ARC-AGI-2 | 8.6% | 11.8% |
| SimpleBench | 58.8% | 46.8% |
| Kagi LLM Benchmark | 74.3% | 78.5% |
| ARC-AGI-1 | 35.7% | 65.3% |
| CritPt | 0.3% | 3.1% |
| EnigmaEval | 5.6% | 3.4% |
| LMArena Hard Prompts | 1399 | 1453 |
| Epoch Capabilities Index | 142.67 | 148.03 |
| NYT Connections (extended) | — | 69.9% |
| Chess Puzzles | — | 12% |
| Thematic Generalization | — | 69.4% |
| DTBench | 81.6% | — |
| LMCA | 37.4% | — |
| ForecastBench | 61.1 | — |
Math Kimi K2.5 leads
Claude Opus 4: 42.0 (#86), Kimi K2.5: 51.8 (#53)
| Benchmark | Claude Opus 4 | Kimi K2.5 |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 64.4% | 92.2% |
| LMArena Math | 1390 | 1470 |
| FrontierMath (Feb 2025 set) | 4.5% | 27.9% |
| FrontierMath Tier 4 (v1) | 4.2% | 4.2% |
| MathArena Final-Answer Competitions | — | 62.3% |
| Omni-MATH | 61.6% | — |
| MATH Level 5 | 85% | — |
Knowledge Kimi K2.5 leads
Claude Opus 4: 44.0 (#88), Kimi K2.5: 53.6 (#56)
| Benchmark | Claude Opus 4 | Kimi K2.5 |
|---|---|---|
| GPQA Diamond | 76.3% | 87.6% |
| Humanity's Last Exam | 10.7% | 24.4% |
| Vectara Hallucination Rate | 12% | 14.2% |
| LMArena Expert | 1386 | 1466 |
| SimpleQA Verified | — | 34.3% |
| MMLU-Pro | 87.5% | — |
| Confabulations | 15.9% | — |
| GPQA (HELM) | 70.8% | — |
Multimodal Kimi K2.5 leads
Claude Opus 4: 31.5 (#106), Kimi K2.5: 41.1 (#39)
| Benchmark | Claude Opus 4 | Kimi K2.5 |
|---|---|---|
| LMArena Vision | 1192 | 1269 |
| GeoBench | 49% | — |
| VPCT | 38% | — |
| LMArena Document | — | 1430 |
Multilingual Kimi K2.5 leads
Claude Opus 4: 48.8 (#138), Kimi K2.5: 53.9 (#53)
| Benchmark | Claude Opus 4 | Kimi K2.5 |
|---|---|---|
| LMArena Non-English | 1362 | 1433 |
| LMArena Chinese | 1386 | 1495 |
| LMArena French | 1372 | 1454 |
| LMArena German | 1391 | 1441 |
| LMArena Japanese | 1331 | 1421 |
| LMArena Korean | 1321 | 1410 |
| LMArena Russian | 1392 | 1435 |
| LMArena Spanish | 1389 | 1450 |
Instruction Following Claude Opus 4 leads
Claude Opus 4: 77.1 (#28), Kimi K2.5: 75.3 (#64)
| Benchmark | Claude Opus 4 | Kimi K2.5 |
|---|---|---|
| LMArena Instruction Following | 1406 | 1431 |
| IFEval | 91.8% | — |
Long Context Kimi K2.5 leads
Claude Opus 4: 39.6 (#172), Kimi K2.5: 52.1 (#7)
| Benchmark | Claude Opus 4 | Kimi K2.5 |
|---|---|---|
| Fiction.LiveBench | 61.1% | 86.1% |
| LMArena Longer Query | 1422 | 1445 |
| CL-bench | — | 19.3% |
| CL-bench Life | — | 13.2% |
Writing & Preference Kimi K2.5 leads
Claude Opus 4: 61.2 (#89), Kimi K2.5: 65.1 (#53)
| Benchmark | Claude Opus 4 | Kimi K2.5 |
|---|---|---|
| LMArena Text | 1377 | 1445 |
| LMArena Creative Writing | 1387 | 1423 |
| EQ-Bench Creative Writing | 1580 | 1579 |
| LMArena Multi-Turn | 1396 | 1444 |
| Short-Story Creative Writing | 83.6% | — |
| WildBench | 85.2% | — |
Frequently asked questions
Is Claude Opus 4 better than Kimi K2.5?
Kimi K2.5 is the stronger model overall, scoring 48.1 to 43.1 on the Noometry Index.
Which is cheaper, Claude Opus 4 or Kimi K2.5?
Kimi K2.5 is cheaper. It lists at $0.45 per million input tokens and $2.25 per million output tokens; Claude Opus 4 lists at $15 and $75.
Is Claude Opus 4 or Kimi K2.5 better for coding?
Kimi K2.5 scores higher on coding benchmarks: 48.8 versus 47.2 in the Noometry coding category.
Which has the bigger context window?
Kimi K2.5 does, with 262K tokens against 200K.
How many benchmarks do Claude Opus 4 and Kimi K2.5 share?
36 benchmarks have published results for both models. Claude Opus 4 has 56 scored results on Noometry and Kimi K2.5 has 51.