# Claude Opus 4.7 vs Claude Sonnet 5

> Claude Opus 4.7 is the stronger model overall, scoring 58.4 to 54.5 on the Noometry Index. Claude Sonnet 5 costs 2.5× less per token, which makes it the better buy when Claude Opus 4.7's lead doesn't matter for your workload.

- Canonical page: https://noometry.com/compare/claude-opus-4-7-vs-claude-sonnet-5
- Last updated: 2026-10-10
- Shared benchmarks: 47

## Summary

- They share 47 benchmarks with published results for both. Claude Opus 4.7 scores higher in 9 categories and Claude Sonnet 5 in 1 category; 8 gaps are clear of the uncertainty.
- The widest gap is in knowledge, where Claude Opus 4.7 leads 62.6 to 55.7.
- The biggest single-benchmark swing is NYT Connections (extended): 39% for Claude Opus 4.7 and 75.1% for Claude Sonnet 5.
- Claude Sonnet 5 is cheaper at $2 / $10 per million input/output tokens, against $5 / $25 for Claude Opus 4.7.

## Snapshot

| | Claude Opus 4.7 | Claude Sonnet 5 |
|---|---|---|
| Provider | Anthropic | Anthropic |
| Noometry Index | 58.4 | 54.5 |
| Rank | 17 | 26 |
| Context | 1M | 1M |
| Input $/M | $5 | $2 |
| Output $/M | $25 | $10 |
| Weights | Proprietary | Proprietary |

## Coding

- Claude Opus 4.7: 59.9 (#10)
- Claude Sonnet 5: 55.5 (#20)

| Benchmark | Claude Opus 4.7 | Claude Sonnet 5 |
|---|---|---|
| FrontierCode | 38.5% | 42.7% |
| LMArena WebDev | 1557 | 1540 |
| SciCode | 54.5% | 53.6% |
| GSO | 44.1% | 37.3% |
| WeirdML | 76.4% | 68.8% |
| LMArena Coding | 1519 | 1483 |
| ALE-Bench | 1,323 | 1,463 |
| SWE-bench Verified | 83.5% | — |
| DeepSWE | — | 53.8% |
| CursorBench | — | 34.1% |
| MirrorCode | 31.1% | — |

## Agentic & Tool Use

- Claude Opus 4.7: 47.9 (#8)
- Claude Sonnet 5: 42.7 (#16)

| Benchmark | Claude Opus 4.7 | Claude Sonnet 5 |
|---|---|---|
| APEX-Agents | 49.2% | 54.5% |
| GBAEval | 43.8% | 65.3% |
| LMArena Search | 1233 | 1194 |
| Vending-Bench 2 | 10,937 | 6,378 |
| Terminal-Bench | 80.2% | — |
| OSWorld 2.0 | 18.2% | — |
| τ²-bench Banking | 40.2% | — |
| PostTrainBench | 28.6% | — |
| ExploitBench | 26.5% | — |
| GDP.pdf | 21% | — |

## Reasoning

- Claude Opus 4.7: 53.7 (#24)
- Claude Sonnet 5: 48.2 (#37)

| Benchmark | Claude Opus 4.7 | Claude Sonnet 5 |
|---|---|---|
| SimpleBench | 61.7% | 60.6% |
| NYT Connections (extended) | 39% | 75.1% |
| CritPt | 12% | 16.9% |
| Chess Puzzles | 30% | 35% |
| LMArena Hard Prompts | 1507 | 1461 |
| Mystery Game Puzzles | 28% | 35% |
| DTBench | 94.7% | 92.5% |
| LMCA | 52.2% | 50% |
| Epoch Capabilities Index | 156.39 | 156.34 |
| ForecastBench | 60.3 | 61.1 |
| ARC-AGI-2 | 75.8% | — |
| Kagi LLM Benchmark | 80.7% | — |
| ARC-AGI-1 | 93.5% | — |
| Thematic Generalization | 72.8% | — |
| EBR-Bench | 19% | — |
| Surface Evolver Bench | — | 60% |
| Bench to the Future 3 | — | 0.14 |

## Math

- Claude Opus 4.7: 66.5 (#20)
- Claude Sonnet 5: 66.1 (#21)

| Benchmark | Claude Opus 4.7 | Claude Sonnet 5 |
|---|---|---|
| FrontierMath (Tiers 1-3) | 70.2% | 65.6% |
| FrontierMath Tier 4 | 31.7% | 29.3% |
| OTIS Mock AIME 2024-2025 | 97.8% | 94.7% |
| ProofBench | 54% | 77% |
| LMArena Math | 1499 | 1467 |
| MathArena Final-Answer Competitions | 73.6% | — |
| FrontierMath (Feb 2025 set) | 43.8% | — |
| FrontierMath Tier 4 (v1) | 22.9% | — |

## Knowledge

- Claude Opus 4.7: 62.6 (#20)
- Claude Sonnet 5: 55.7 (#42)

| Benchmark | Claude Opus 4.7 | Claude Sonnet 5 |
|---|---|---|
| GPQA Diamond | 90.2% | 90.5% |
| SimpleQA Verified | 51.7% | 33.7% |
| LMArena Expert | 1521 | 1491 |
| Humanity's Last Exam | 36.2% | — |
| Vectara Hallucination Rate | 12% | — |

## Multimodal

- Claude Opus 4.7: 41.5 (#29)
- Claude Sonnet 5: 42.4 (#26)

| Benchmark | Claude Opus 4.7 | Claude Sonnet 5 |
|---|---|---|
| LMArena Vision | 1317 | 1275 |
| Blueprint-Bench 2 | 24.5% | 24.9% |
| LMArena Document | 1495 | 1466 |
| Furniture Assembly | 33.3% | — |

## Multilingual

- Claude Opus 4.7: 57.5 (#9)
- Claude Sonnet 5: 54.0 (#55)

| Benchmark | Claude Opus 4.7 | Claude Sonnet 5 |
|---|---|---|
| LMArena Non-English | 1481 | 1431 |
| LMArena Chinese | 1530 | 1475 |
| LMArena French | 1505 | 1457 |
| LMArena German | 1494 | 1448 |
| LMArena Japanese | 1472 | 1422 |
| LMArena Korean | 1466 | 1411 |
| LMArena Russian | 1494 | 1452 |
| LMArena Spanish | 1497 | 1436 |

## Instruction Following

- Claude Opus 4.7: 78.6 (#9)
- Claude Sonnet 5: 76.6 (#39)

| Benchmark | Claude Opus 4.7 | Claude Sonnet 5 |
|---|---|---|
| LMArena Instruction Following | 1498 | 1453 |

## Long Context

- Claude Opus 4.7: 46.1 (#24)
- Claude Sonnet 5: 44.8 (#53)

| Benchmark | Claude Opus 4.7 | Claude Sonnet 5 |
|---|---|---|
| LMArena Longer Query | 1505 | 1464 |

## Writing & Preference

- Claude Opus 4.7: 75.3 (#8)
- Claude Sonnet 5: 69.4 (#24)

| Benchmark | Claude Opus 4.7 | Claude Sonnet 5 |
|---|---|---|
| LMArena Text | 1491 | 1443 |
| LMArena Creative Writing | 1486 | 1416 |
| EQ-Bench Creative Writing | 1914 | 1794 |
| EQ-Bench 4 | 1311 | 1236 |
| LMArena Multi-Turn | 1506 | 1454 |

## FAQ

### Is Claude Opus 4.7 better than Claude Sonnet 5?

Claude Opus 4.7 is the stronger model overall, scoring 58.4 to 54.5 on the Noometry Index. Claude Sonnet 5 costs 2.5× less per token, which makes it the better buy when Claude Opus 4.7's lead doesn't matter for your workload.

### Which is cheaper, Claude Opus 4.7 or Claude Sonnet 5?

Claude Sonnet 5 is cheaper. It lists at $2 per million input tokens and $10 per million output tokens; Claude Opus 4.7 lists at $5 and $25.

### Is Claude Opus 4.7 or Claude Sonnet 5 better for coding?

Claude Opus 4.7 scores higher on coding benchmarks: 59.9 versus 55.5 in the Noometry coding category.

### Which has the bigger context window?

Both accept 1M tokens.

### How many benchmarks do Claude Opus 4.7 and Claude Sonnet 5 share?

47 benchmarks have published results for both models. Claude Opus 4.7 has 66 scored results on Noometry and Claude Sonnet 5 has 51.