# Claude Sonnet 4.6 vs Gemini 2.5 Pro

> Claude Sonnet 4.6 is the stronger model overall, scoring 50.1 to 44.9 on the Noometry Index. Gemini 2.5 Pro costs 1.7× less per token, which makes it the better buy when Claude Sonnet 4.6's lead doesn't matter for your workload.

- Canonical page: https://noometry.com/compare/claude-sonnet-4-6-vs-gemini-2-5-pro
- Last updated: 2026-10-10
- Shared benchmarks: 42

## Summary

- They share 42 benchmarks with published results for both. Claude Sonnet 4.6 scores higher in 6 categories and Gemini 2.5 Pro in 4 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in math, where Claude Sonnet 4.6 leads 51.9 to 31.9.
- The biggest single-benchmark swing is ARC-AGI-2: 60.4% for Claude Sonnet 4.6 and 4.9% for Gemini 2.5 Pro.
- Gemini 2.5 Pro is cheaper at $1.25 / $10 per million input/output tokens, against $3 / $15 for Claude Sonnet 4.6.
- Gemini 2.5 Pro accepts more context: 1.05M tokens versus 1M.

## Snapshot

| | Claude Sonnet 4.6 | Gemini 2.5 Pro |
|---|---|---|
| Provider | Anthropic | Google |
| Noometry Index | 50.1 | 44.9 |
| Rank | 42 | 67 |
| Context | 1M | 1.05M |
| Input $/M | $3 | $1.25 |
| Output $/M | $15 | $10 |
| Weights | Proprietary | Proprietary |

## Coding

- Claude Sonnet 4.6: 46.0 (#62)
- Gemini 2.5 Pro: 42.8 (#86)

| Benchmark | Claude Sonnet 4.6 | Gemini 2.5 Pro |
|---|---|---|
| SWE-bench Verified | 75.2% | 57.6% |
| LMArena WebDev | 1521 | 1227 |
| SciCode | 46.8% | 42.8% |
| WeirdML | 66.1% | 54% |
| LMArena Coding | 1504 | 1452 |
| ALE-Bench | 1,327 | 785.52 |
| DeepSWE | 29.9% | — |
| FrontierCode | 24.3% | — |
| SWE-bench Verified (bash only) | — | 53.6% |
| Aider Polyglot | — | 83.1% |
| GSO | — | 3.9% |
| LiveBench Coding | — | 85.9% |
| CadEval | — | 64% |
| AlgoTune | — | 1.51 |

## Agentic & Tool Use

- Claude Sonnet 4.6: 39.1 (#26)
- Gemini 2.5 Pro: 29.1 (#81)

| Benchmark | Claude Sonnet 4.6 | Gemini 2.5 Pro |
|---|---|---|
| Terminal-Bench | 53.4% | 32.6% |
| DeepResearch Bench | 54.9% | 42.8% |
| LMArena Search | 1221 | 1142 |
| Vending-Bench 2 | 7,204 | 573.64 |
| APEX-Agents | 43% | — |
| OSWorld 2.0 | 9.3% | — |
| GDPval | — | 23.3% |
| Remote Labor Index | — | 0.8% |
| TheAgentCompany | — | 30.3% |
| τ²-bench Banking | — | 13.7% |
| OSWorld | 72.1% | — |
| BALROG | — | 43.3% |
| ExploitBench | 23.6% | — |
| GBAEval | 48.8% | — |
| GDP.pdf | 18% | — |
| METR Time Horizons | — | 55.4% |

## Reasoning

- Claude Sonnet 4.6: 46.1 (#40)
- Gemini 2.5 Pro: 28.8 (#85)

| Benchmark | Claude Sonnet 4.6 | Gemini 2.5 Pro |
|---|---|---|
| ARC-AGI-2 | 60.4% | 4.9% |
| ARC-AGI-1 | 86.5% | 41% |
| CritPt | 3.1% | 2% |
| Chess Puzzles | 13% | 20% |
| LMArena Hard Prompts | 1485 | 1455 |
| DTBench | 89.9% | 82.4% |
| LMCA | 46.5% | 34.8% |
| Epoch Capabilities Index | 152.25 | 145.32 |
| ForecastBench | 62 | 61.3 |
| SimpleBench | — | 62.4% |
| Kagi LLM Benchmark | — | 70.3% |
| NYT Connections (extended) | 80.9% | — |
| EnigmaEval | — | 5.6% |
| Thematic Generalization | 76.3% | — |
| LiveBench Reasoning | — | 89.8% |
| Mystery Game Puzzles | 16% | — |
| LiveBench Data Analysis | — | 79.9% |
| LiveBench | — | 82.3% |

## Math

- Claude Sonnet 4.6: 51.9 (#41)
- Gemini 2.5 Pro: 31.9 (#191)

| Benchmark | Claude Sonnet 4.6 | Gemini 2.5 Pro |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 85.8% | 84.7% |
| LMArena Math | 1463 | 1450 |
| FrontierMath (Feb 2025 set) | 32.4% | 14.1% |
| FrontierMath Tier 4 (v1) | 8.3% | 4.2% |
| FrontierMath (Tiers 1-3) | — | 24.6% |
| FrontierMath Tier 4 | — | 0% |
| ProofBench | 45% | — |
| Omni-MATH | — | 41.6% |
| LiveBench Math | — | 90.2% |
| MATH Level 5 | — | 95.9% |

## Knowledge

- Claude Sonnet 4.6: 51.6 (#60)
- Gemini 2.5 Pro: 55.8 (#41)

| Benchmark | Claude Sonnet 4.6 | Gemini 2.5 Pro |
|---|---|---|
| GPQA Diamond | 87.4% | 85.3% |
| Vectara Hallucination Rate | 10.6% | 7% |
| LMArena Expert | 1500 | 1452 |
| Humanity's Last Exam | — | 21.6% |
| SimpleQA Verified | 35.5% | — |
| MMLU-Pro | — | 86.3% |
| Confabulations | — | 10.6% |
| GPQA (HELM) | — | 74.9% |

## Multimodal

- Claude Sonnet 4.6: 38.0 (#61)
- Gemini 2.5 Pro: 44.5 (#16)

| Benchmark | Claude Sonnet 4.6 | Gemini 2.5 Pro |
|---|---|---|
| LMArena Vision | 1283 | 1263 |
| LMArena Document | 1482 | 1421 |
| GeoBench | — | 86% |
| VPCT | — | 48% |
| Blueprint-Bench 2 | 6.7% | — |
| SpatialViz-Bench | — | 44.7% |

## Multilingual

- Claude Sonnet 4.6: 54.6 (#39)
- Gemini 2.5 Pro: 55.5 (#30)

| Benchmark | Claude Sonnet 4.6 | Gemini 2.5 Pro |
|---|---|---|
| LMArena Non-English | 1440 | 1452 |
| LMArena Chinese | 1491 | 1507 |
| LMArena French | 1465 | 1473 |
| LMArena German | 1428 | 1487 |
| LMArena Japanese | 1419 | 1462 |
| LMArena Korean | 1411 | 1435 |
| LMArena Russian | 1440 | 1460 |
| LMArena Spanish | 1465 | 1473 |

## Instruction Following

- Claude Sonnet 4.6: 77.6 (#23)
- Gemini 2.5 Pro: 75.2 (#71)

| Benchmark | Claude Sonnet 4.6 | Gemini 2.5 Pro |
|---|---|---|
| LMArena Instruction Following | 1476 | 1436 |
| LiveBench Instruction Following | — | 80.6% |
| IFEval | — | 84% |

## Long Context

- Claude Sonnet 4.6: 45.3 (#40)
- Gemini 2.5 Pro: 59.7 (#5)

| Benchmark | Claude Sonnet 4.6 | Gemini 2.5 Pro |
|---|---|---|
| LMArena Longer Query | 1479 | 1449 |
| Fiction.LiveBench | — | 91.7% |

## Writing & Preference

- Claude Sonnet 4.6: 70.4 (#22)
- Gemini 2.5 Pro: 63.9 (#60)

| Benchmark | Claude Sonnet 4.6 | Gemini 2.5 Pro |
|---|---|---|
| LMArena Text | 1458 | 1458 |
| LMArena Creative Writing | 1435 | 1453 |
| EQ-Bench Creative Writing | 1810 | 1421 |
| LMArena Multi-Turn | 1464 | 1453 |
| Short-Story Creative Writing | — | 83.8% |
| WildBench | — | 85.7% |
| EQ-Bench 4 | 1207 | — |
| LiveBench Language | — | 67.8% |

## FAQ

### Is Claude Sonnet 4.6 better than Gemini 2.5 Pro?

Claude Sonnet 4.6 is the stronger model overall, scoring 50.1 to 44.9 on the Noometry Index. Gemini 2.5 Pro costs 1.7× less per token, which makes it the better buy when Claude Sonnet 4.6's lead doesn't matter for your workload.

### Which is cheaper, Claude Sonnet 4.6 or Gemini 2.5 Pro?

Gemini 2.5 Pro is cheaper. It lists at $1.25 per million input tokens and $10 per million output tokens; Claude Sonnet 4.6 lists at $3 and $15.

### Is Claude Sonnet 4.6 or Gemini 2.5 Pro better for coding?

Claude Sonnet 4.6 scores higher on coding benchmarks: 46.0 versus 42.8 in the Noometry coding category.

### Which has the bigger context window?

Gemini 2.5 Pro does, with 1.05M tokens against 1M.

### How many benchmarks do Claude Sonnet 4.6 and Gemini 2.5 Pro share?

42 benchmarks have published results for both models. Claude Sonnet 4.6 has 57 scored results on Noometry and Gemini 2.5 Pro has 78.