# Claude Opus 5.5 vs Claude Sonnet 5.5

> Side-by-side benchmark comparison for Claude Opus 5.5 and Claude Sonnet 5.5 across agentic, coding, multimodal, knowledge, reasoning, multilingual, and math tasks.

- Canonical page: https://benchlm.ai/compare/claude-opus-5-5-vs-claude-sonnet-5-5
- Last updated: September 28, 2026

- Shared sourced benchmarks: 45
- HTML indexing: indexable
- Ranking lane: BenchAlign v5.7

## Quick Verdict

Pick Claude Opus 5.5 if you want the stronger benchmark profile. Claude Sonnet 5.5 only makes more sense when its price, context window, or workload-specific category wins matter more than the overall score.

## Summary

- Claude Opus 5.5 leads overall 87.07 to 80.49.
- The clearest category separation is in agentic, where the averages are 87.8 for Claude Opus 5.5 and 66.1 for Claude Sonnet 5.5.
- The biggest single benchmark swing is ProgramBench in Coding, with scores of 91.2% and 79.7%.
- Claude Sonnet 5.5 is the cheaper option on output tokens, which matters if you expect large responses or heavy interactive use.

## Model Snapshot

| Property | Claude Opus 5.5 | Claude Sonnet 5.5 |
|----------|----------|----------|
| Creator | Anthropic | Anthropic |
| Type | Proprietary | Proprietary |
| Reasoning | Reasoning | Reasoning |
| Context | 1M | 1M |
| Overall Score | 87.07 | 80.49 |
| Benchmarks Covered | 49 | 47 |
| Pricing (input/output) | $4.00 / $20.00 | $2.00 / $10.00 |

## Category Breakdown

### Agentic

- Winner: Claude Opus 5.5
- Claude Opus 5.5 public-lane score: 87.8 (Supported · #1/111)
- Claude Sonnet 5.5 public-lane score: 66.1 (Supported · #10/111)

| Benchmark | Claude Opus 5.5 | Claude Sonnet 5.5 | Winner |
|-----------|-----------|-----------|--------|
| Terminal-Bench 4.0 | 66.40% | 70.60% | Claude Sonnet 5.5 |
| Terminal-Bench-Science 0.1 | 58.7% | 59.9% | Claude Sonnet 5.5 |
| AutomationBench | 40.0% | Coming soon | Coming soon |
| HLE w/ tools | 67.7% | 64.5% | Claude Opus 5.5 |
| OSWorld 2.0 | 48.7% | Coming soon | Coming soon |
| LAB all-pass (Harvey held-out) | 8.3% | 10.0% | Claude Sonnet 5.5 |
| LAB criterion-pass (Harvey held-out) | 91.2% | 93.1% | Claude Sonnet 5.5 |
| Toolathlon-Verified | 77.8% | 77.8% | Tie |
| Toolathlon Verified Pass@3 | 82.4% | 85.2% | Claude Sonnet 5.5 |
| Toolathlon Verified Pass³ | 72.2% | 68.5% | Claude Opus 5.5 |
| Toolathlon Verified avg. turns | 26.9 turns | 31.6 turns | Claude Sonnet 5.5 |
| DRACO | Coming soon | 87.0% | Coming soon |
| AutomationBench (Zapier 1.0.6) | Coming soon | 44.7% | Coming soon |

### Coding

- Winner: Directional only
- Claude Opus 5.5 public-lane score: 83 (Supported · #1/136)
- Claude Sonnet 5.5 public-lane score: 79.6 (Estimated · #3/136)

| Benchmark | Claude Opus 5.5 | Claude Sonnet 5.5 | Winner |
|-----------|-----------|-----------|--------|
| FrontierCode 1.1 Main | 54.4% | 46.2% | Claude Opus 5.5 |
| cursorBench40 | 57.8% | 55.5% | Claude Opus 5.5 |
| SWE-bench Pro | 89.9% | 81.3% | Claude Opus 5.5 |
| SWE Multilingual | 93.9% | 90.3% | Claude Opus 5.5 |
| SWE Multimodal | 61.4% | 54.3% | Claude Opus 5.5 |
| DeepSWE | 74.2% | 71.0% | Claude Opus 5.5 |
| FrontierCode 1.1 Extended | 63.6% | 59.1% | Claude Opus 5.5 |
| FrontierSWE v2 | 62.3% | 61.9% | Claude Opus 5.5 |
| ProgramBench | 91.2% | 79.7% | Claude Opus 5.5 |

### Reasoning

- Winner: Directional only
- Claude Opus 5.5 public-lane score: 82.4 (#2/27)
- Claude Sonnet 5.5 public-lane score: 79.2 (#7/27)

| Benchmark | Claude Opus 5.5 | Claude Sonnet 5.5 | Winner |
|-----------|-----------|-----------|--------|
| ARC-AGI-1 | 97.50% | Coming soon | Coming soon |
| ARC-AGI-2 | 91.7% | Coming soon | Coming soon |

### Multimodal & Grounded

- Winner: Not comparable
- Claude Opus 5.5 public-lane score: 88.8 (#3/50)
- Claude Sonnet 5.5 public-lane score: 83.3 (Unranked · 7 rankable rows)

| Benchmark | Claude Opus 5.5 | Claude Sonnet 5.5 | Winner |
|-----------|-----------|-----------|--------|
| Chartography (tools) | 89.0% | 90.2% | Claude Sonnet 5.5 |
| Chartography (no tools) | 64.4% | 61.6% | Claude Opus 5.5 |
| BenchCAD Vision2Code (no tools) | 0.730 | 0.747 | Claude Sonnet 5.5 |
| BenchCAD Vision2Code (tools) | 0.962 | 0.963 | Claude Sonnet 5.5 |
| Biomedical image analysis | 71.4% | 72.2% | Claude Sonnet 5.5 |
| OfficeQA | 78.9% | 76.9% | Claude Opus 5.5 |
| OfficeQA Pro | 67.7% | 65.6% | Claude Opus 5.5 |

### Knowledge

- Winner: Claude Opus 5.5
- Claude Opus 5.5 public-lane score: 89.1 (Supported · #1/160)
- Claude Sonnet 5.5 public-lane score: 80.9 (Supported · #6/160)

| Benchmark | Claude Opus 5.5 | Claude Sonnet 5.5 | Winner |
|-----------|-----------|-----------|--------|
| HLE w/o tools | 64.4% | 56.9% | Claude Opus 5.5 |
| HealthBench (raw) | 68.1% | 69.4% | Claude Sonnet 5.5 |
| HealthBench (length-adjusted) | 60.6% | 65.4% | Claude Sonnet 5.5 |
| HealthBench Professional | 65.6% | 69.2% | Claude Sonnet 5.5 |
| HealthBench Professional (raw) | 77.1% | 77.1% | Tie |
| BioMysteryBench (human-solvable) | 89.3% | 89.2% | Claude Opus 5.5 |
| BioMysteryBench (human-difficult) | 50.0% | 44.7% | Claude Opus 5.5 |
| SpatialBench Verified | 72.0% | 72.5% | Claude Sonnet 5.5 |
| SingleCellBench | 61.2% | 59.1% | Claude Opus 5.5 |
| Morphology-to-molecule matching | 34.0% | 25.0% | Claude Opus 5.5 |
| Medicinal chemistry | 63.5% | 65.3% | Claude Sonnet 5.5 |
| Protein Design | 60.2% | 51.0% | Claude Opus 5.5 |
| Protein Design library ranking | 56.0% | 54.8% | Claude Opus 5.5 |
| De novo protein-binder design | 82.6% | 82.3% | Claude Opus 5.5 |
| Protocols (troubleshooting) | 73.7% | 67.3% | Claude Opus 5.5 |
| Protocols (understanding) | 69.0% | 66.6% | Claude Opus 5.5 |

### Multilingual

- Winner: Not comparable
- Claude Opus 5.5 public-lane score: Coming soon
- Claude Sonnet 5.5 public-lane score: Coming soon

| Benchmark | Claude Opus 5.5 | Claude Sonnet 5.5 | Winner |
|-----------|-----------|-----------|--------|
| GMMLU | 94.3% | 92.1% | Claude Opus 5.5 |
| MILU | 93.1% | 91.6% | Claude Opus 5.5 |

### Mathematics

- Winner: Not comparable
- Claude Opus 5.5 public-lane score: Coming soon
- Claude Sonnet 5.5 public-lane score: Coming soon

| Benchmark | Claude Opus 5.5 | Claude Sonnet 5.5 | Winner |
|-----------|-----------|-----------|--------|
| ArXivMath Aug. 2026 (no tools) | 91.2% | 86.8% | Claude Opus 5.5 |
| ArXivMath Aug. 2026 (tools) | 96.9% | 95.2% | Claude Opus 5.5 |

## FAQ

### Which is better overall, Claude Opus 5.5 or Claude Sonnet 5.5?

Claude Opus 5.5 is ahead overall on BenchLM right now.

### Where is the biggest gap between Claude Opus 5.5 and Claude Sonnet 5.5?

The widest category gap is in agentic, where the averages are 87.8 for Claude Opus 5.5 and 66.1 for Claude Sonnet 5.5.

### How many benchmarks does Claude Opus 5.5 cover on BenchLM?

Claude Opus 5.5 currently has 49 sourced benchmark scores on BenchLM.

### How many benchmarks does Claude Sonnet 5.5 cover on BenchLM?

Claude Sonnet 5.5 currently has 47 sourced benchmark scores on BenchLM.

## Related Comparisons

- [Claude Opus 5.5 vs GPT-6 Astra](/compare/claude-opus-5-5-vs-gpt-6-astra)
- [Claude Sonnet 5.5 vs GPT-6 Astra](/compare/claude-sonnet-5-5-vs-gpt-6-astra)
- [Claude Opus 5.5 vs Claude Fable 5.1](/compare/claude-fable-5-1-vs-claude-opus-5-5)
- [Claude Sonnet 5.5 vs Claude Fable 5.1](/compare/claude-fable-5-1-vs-claude-sonnet-5-5)

## Explore More

- [Claude Opus 5.5 profile](/models/claude-opus-5-5)
- [Claude Sonnet 5.5 profile](/models/claude-sonnet-5-5)
- [Compare Pricing](/llm-pricing)
- [Alternative Finder](/tools/alternative-finder)
- [LLM Selector](/tools/llm-selector)
- [Overall Rankings](/best/overall)