Claude Opus 5.5 vs o3
Claude Opus 5.5 
o3 Verdict
- Claude Opus 5.5 (thinking) leads on quality: 70.7 vs 60.1.
- Claude Opus 5.5 (thinking) is stronger in composite, knowledge, long context, multimodal, reasoning.
- o3 is stronger in agents & tools, coding, human preference, instruction following, maths.
- o3 is 2.3× cheaper ($3.50 vs $8.00 per 1M blended).
- o3 streams 2.6× faster (124 vs 48 tokens per second).
| Metric | Claude Opus 5.5 (thinking) | o3 |
|---|---|---|
| BenchLeader Index | 70.7 | 60.1 |
| Composite score | 95.0 | 53.8 |
| Knowledge score | 85.0 | 56.9 |
| Long context score | 68.5 | 63.4 |
| Multimodal score | 71.9 | 57.7 |
| Reasoning score | 95.0 | 52.7 |
| Agents & tools score | – | 70.2 |
| Coding score | – | 61.9 |
| Human preference score | – | 62.3 |
| Instruction following score | – | 65.1 |
| Maths score | – | 71.2 |
| Blended price $/M | $8.00 | $3.50 |
| Output speed | 48 tok/s | 124 tok/s |
| Time to first answer | 4.2 s | 7.7 s |
| Context window | 1M | 200k |
| Epoch Capabilities Index | – | 146.9 |
| LMArena Text | – | 1432 |
| LMArena Hard Prompts | – | 1441 |
| LMArena Coding | – | 1460 |
| LMArena Vision | – | 1214 |
| AA Intelligence Index | 57.6 | 20.2 |
| IFBench | – | 71.4% |
| AA-LCR | 84.7% | 74.7% |
| MMMU-Pro | 87.7% | 70.1% |
| AA-Omniscience | 46.4 | -15.6 |
| Terminal-Bench Hard | – | 37.1% |
| GPQA Diamond (AA) | – | 82.7% |
| Humanity's Last Exam (AA) | 61.4% | 20.1% |
| SciCode (AA) | 66.9% | – |
| τ²-Bench Telecom (AA) | – | 80.7% |
| PRBench Finance | – | 47.7% |
| PRBench Legal | – | 48.6% |
| MMMU (validation) | – | 82.9% |
| MMMU-Pro (official) | – | 76.4% |
| Kagi LLM Benchmark | – | 67.6% |
| IFEval (HELM) | – | 86.9% |
| Omni-MATH (HELM) | – | 71.4% |
| WildBench (HELM) | – | 86.1% |
| MMLU-Pro (HELM) | – | 85.9% |
| GPQA Diamond (HELM) | – | 75.3% |
| HELM Capabilities mean | – | 81.1% |
| Aider Polyglot | – | 81.3% |
| SWE-bench Verified (bash only) | – | 58.4% |
| SWE-bench Verified (any scaffold) | – | 58.4% |
| BFCL Overall | – | 63.0% |
| CritPt | 31.7% | 1.1% |
| GDPval (AA) | 67.3% | – |
| FORTRESS | – | 16.0% |
| PropensityBench | – | 10.5% |
| SciPredict | – | 17.9% |
| VTB | – | 13.7% |
| LMArena Maths | – | 1448 |
| LMArena Creative Writing | – | 1382 |
| LMArena Instruction Following | – | 1402 |
| LMArena Multi-turn | – | 1420 |
| LMArena Longer Queries | – | 1410 |
| METR Time Horizons | – | 63.6% |
| ForecastBench | – | 62.5% |
| Terminal-Bench 4.0 (AA) | 59.6% | – |
| AutomationBench | 69.5% | – |
| GDP.pdf | 26.2% | – |
| Harvey LAB | 91.2% | – |
| AA-Omniscience: accuracy | 66.2% | 38.6% |
| AA-Omniscience: non-hallucination | 41.4% | 11.9% |
| AA-Briefcase | 1822 | – |
| BrowseComp-Plus | – | 50.5% |
Data as of 2026-09-22. Best configuration of each model; every score links to its source on the model pages.
Claude Opus 5.5 vs o3: questions
- Is Claude Opus 5.5 better than o3?
- Claude Opus 5.5 (thinking) leads on quality: 70.7 vs 60.1. The BenchLeader Index combines every independent quality benchmark; Claude Opus 5.5 (thinking) is ahead overall as of 2026-09-22, but check the category scores for your use.
- Which is cheaper, Claude Opus 5.5 or o3?
- o3 is cheaper: $3.50 against $8.00 per million tokens, blended at three input tokens per output token.
- Which is faster, Claude Opus 5.5 or o3?
- o3 streams faster: 124 against 48 output tokens per second.
- Which has the larger context window?
- Claude Opus 5.5 accepts more context: 1M against 200k tokens.