# Claude Opus 4.5 vs Kimi K3

> Kimi K3 is the stronger model overall, scoring 59.5 to 50.5 on the Noometry Index.

- Canonical page: https://noometry.com/compare/claude-opus-4-5-vs-kimi-k3
- Last updated: 2026-10-10
- Shared benchmarks: 40

## Summary

- They share 40 benchmarks with published results for both. Claude Opus 4.5 scores higher in 2 categories and Kimi K3 in 8 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in math, where Kimi K3 leads 74.0 to 38.5.
- The biggest single-benchmark swing is ProofBench: 36% for Claude Opus 4.5 and 87% for Kimi K3.
- Kimi K3 is cheaper at $3 / $15 per million input/output tokens, against $5 / $25 for Claude Opus 4.5.
- Kimi K3 accepts more context: 1.05M tokens versus 200K.
- Kimi K3 has downloadable open weights; the other is API-only.

## Snapshot

| | Claude Opus 4.5 | Kimi K3 |
|---|---|---|
| Provider | Anthropic | Moonshot AI |
| Noometry Index | 50.5 | 59.5 |
| Rank | 39 | 12 |
| Context | 200K | 1.05M |
| Input $/M | $5 | $3 |
| Output $/M | $25 | $15 |
| Weights | Proprietary | Open |

## Coding

- Claude Opus 4.5: 55.4 (#21)
- Kimi K3: 60.6 (#8)

| Benchmark | Claude Opus 4.5 | Kimi K3 |
|---|---|---|
| LMArena WebDev | 1493 | 1659 |
| WeirdML | 63.7% | 82.6% |
| LMArena Coding | 1504 | 1509 |
| ALE-Bench | 1,025 | 1,524 |
| SWE-bench Verified | 76.7% | — |
| DeepSWE | — | 68.5% |
| FrontierCode | — | 44.2% |
| SWE-bench Verified (bash only) | 76.8% | — |
| SWE-bench Multilingual | 70.7% | — |
| FrontierSWE | — | 25.9% |
| SciCode | — | 59.5% |
| GSO | 26.5% | — |
| AlgoTune | 1.77 | — |

## Agentic & Tool Use

- Claude Opus 4.5: 47.2 (#11)
- Kimi K3: 41.8 (#19)

| Benchmark | Claude Opus 4.5 | Kimi K3 |
|---|---|---|
| τ²-bench Banking | 24.7% | 37.1% |
| Vending-Bench 2 | 4,967 | 5,165 |
| Terminal-Bench | 63.1% | — |
| APEX-Agents | — | 50.6% |
| Berkeley Function Calling Leaderboard | 77.5% | — |
| GDPval | 45.5% | — |
| Remote Labor Index | 3.8% | — |
| τ²-bench Airline | 84% | — |
| τ²-bench Retail | 79.6% | — |
| τ²-bench Telecom | 92.3% | — |
| Cybench | 82% | — |
| DeepResearch Bench | 54.8% | — |
| OSWorld | 66.3% | — |
| PostTrainBench | — | 32% |
| BALROG | 43.5% | — |
| GBAEval | — | 48.3% |
| GDP.pdf | — | 19% |
| LMArena Search | 1180 | — |
| METR Time Horizons | 75% | — |

## Reasoning

- Claude Opus 4.5: 42.5 (#43)
- Kimi K3: 63.0 (#15)

| Benchmark | Claude Opus 4.5 | Kimi K3 |
|---|---|---|
| ARC-AGI-2 | 37.6% | 60.4% |
| SimpleBench | 62% | 60.7% |
| NYT Connections (extended) | 52.5% | 93.6% |
| ARC-AGI-1 | 80% | 94.5% |
| Chess Puzzles | 12% | 39% |
| LMArena Hard Prompts | 1476 | 1495 |
| Mystery Game Puzzles | 22% | 26% |
| DTBench | 89.9% | 91.2% |
| LMCA | 44.5% | 52.7% |
| Epoch Capabilities Index | 150.1 | 157.68 |
| ForecastBench | 60.7 | 61.1 |
| Kagi LLM Benchmark | 80.2% | — |
| CritPt | — | 23.4% |
| EnigmaEval | 11.9% | — |
| EBR-Bench | 14.3% | — |
| Surface Evolver Bench | — | 95% |

## Math

- Claude Opus 4.5: 38.5 (#85)
- Kimi K3: 74.0 (#13)

| Benchmark | Claude Opus 4.5 | Kimi K3 |
|---|---|---|
| FrontierMath (Tiers 1-3) | 34.4% | 72.2% |
| FrontierMath Tier 4 | 4.9% | 39% |
| OTIS Mock AIME 2024-2025 | 86.1% | 97.2% |
| ProofBench | 36% | 87% |
| LMArena Math | 1463 | 1493 |
| MathArena Final-Answer Competitions | — | 87.8% |
| FrontierMath (Feb 2025 set) | 20.7% | — |
| FrontierMath Tier 4 (v1) | 4.2% | — |

## Knowledge

- Claude Opus 4.5: 56.4 (#38)
- Kimi K3: 63.2 (#19)

| Benchmark | Claude Opus 4.5 | Kimi K3 |
|---|---|---|
| GPQA Diamond | 86% | 93.1% |
| SimpleQA Verified | 45.7% | 50.6% |
| LMArena Expert | 1488 | 1520 |
| Humanity's Last Exam | 25.2% | — |
| Vectara Hallucination Rate | 10.9% | — |

## Multimodal

- Claude Opus 4.5: 31.1 (#99)
- Kimi K3: 38.4 (#58)

| Benchmark | Claude Opus 4.5 | Kimi K3 |
|---|---|---|
| Furniture Assembly | 28.3% | 34.2% |
| GeoBench | 75% | — |
| VPCT | 40% | — |
| Blueprint-Bench 2 | — | 29.5% |
| LMArena Document | 1462 | — |

## Multilingual

- Claude Opus 4.5: 54.5 (#45)
- Kimi K3: 56.5 (#20)

| Benchmark | Claude Opus 4.5 | Kimi K3 |
|---|---|---|
| LMArena Non-English | 1438 | 1466 |
| LMArena Chinese | 1470 | 1528 |
| LMArena French | 1471 | 1488 |
| LMArena German | 1449 | 1474 |
| LMArena Japanese | 1415 | 1492 |
| LMArena Korean | 1424 | 1465 |
| LMArena Russian | 1448 | 1481 |
| LMArena Spanish | 1455 | 1478 |

## Instruction Following

- Claude Opus 4.5: 77.7 (#20)
- Kimi K3: 77.8 (#13)

| Benchmark | Claude Opus 4.5 | Kimi K3 |
|---|---|---|
| LMArena Instruction Following | 1478 | 1481 |

## Long Context

- Claude Opus 4.5: 46.4 (#21)
- Kimi K3: 45.8 (#27)

| Benchmark | Claude Opus 4.5 | Kimi K3 |
|---|---|---|
| LMArena Longer Query | 1480 | 1494 |
| CL-bench | 21.1% | — |

## Writing & Preference

- Claude Opus 4.5: 68.2 (#28)
- Kimi K3: 76.7 (#4)

| Benchmark | Claude Opus 4.5 | Kimi K3 |
|---|---|---|
| LMArena Text | 1451 | 1476 |
| LMArena Creative Writing | 1445 | 1452 |
| EQ-Bench Creative Writing | 1687 | 2082 |
| LMArena Multi-Turn | 1466 | 1488 |
| EQ-Bench 4 | — | 1339 |

## FAQ

### Is Claude Opus 4.5 better than Kimi K3?

Kimi K3 is the stronger model overall, scoring 59.5 to 50.5 on the Noometry Index.

### Which is cheaper, Claude Opus 4.5 or Kimi K3?

Kimi K3 is cheaper. It lists at $3 per million input tokens and $15 per million output tokens; Claude Opus 4.5 lists at $5 and $25.

### Is Claude Opus 4.5 or Kimi K3 better for coding?

Kimi K3 scores higher on coding benchmarks: 60.6 versus 55.4 in the Noometry coding category.

### Which has the bigger context window?

Kimi K3 does, with 1.05M tokens against 200K.

### How many benchmarks do Claude Opus 4.5 and Kimi K3 share?

40 benchmarks have published results for both models. Claude Opus 4.5 has 69 scored results on Noometry and Kimi K3 has 53.