Claude Sonnet 5.5 vs o3
Claude Sonnet 5.5 
o3 Verdict
- Claude Sonnet 5.5 (max) leads on quality: 67.6 vs 59.7.
- Claude Sonnet 5.5 (max) is stronger in coding, composite, knowledge, long context, maths, reasoning.
- o3 is stronger in agents & tools, human preference, instruction following, multimodal.
- They cost about the same ($3.50 per 1M blended).
| Metric | Claude Sonnet 5.5 (max) | o3 |
|---|---|---|
| BenchLeader Index | 67.6 | 59.7 |
| Agents & tools score | 69.5 | 70.2 |
| Coding score | 69.8 | 61.8 |
| Composite score | 72.6 | 52.3 |
| Knowledge score | 65.4 | 56.2 |
| Long context score | 66.6 | 62.5 |
| Maths score | 72.9 | 70.8 |
| Reasoning score | 81.1 | 52.1 |
| Human preference score | – | 61.7 |
| Instruction following score | – | 65.1 |
| Multimodal score | – | 57.0 |
| Blended price $/M | $4.00 | $3.50 |
| Output speed | 141 tok/s | 116 tok/s |
| Time to first answer | 487.1 s | 10.7 s |
| Context window | 1M | 200k |
| GPQA Diamond | 95.6% | – |
| FrontierMath Tiers 1–3 | 88.8% | – |
| FrontierMath Tier 4 | 80.5% | – |
| OTIS Mock AIME | 100.0% | – |
| SimpleQA Verified | 46.5% | – |
| Terminal-Bench | 61.8% | – |
| SciCode | 61.0% | – |
| APEX-Agents | 75.5% | – |
| ProofBench | 100.0% | – |
| Epoch Capabilities Index | – | 146.9 |
| LMArena Text | – | 1432 |
| LMArena Hard Prompts | – | 1442 |
| LMArena Coding | – | 1460 |
| LMArena Vision | – | 1214 |
| LMArena Agent | 12 | – |
| LiveBench | 75.7% | – |
| LiveBench Reasoning | 91.6% | – |
| LiveBench Coding | 91.4% | – |
| LiveBench Agentic Coding | 56.3% | – |
| LiveBench Mathematics | 96.1% | – |
| LiveBench Data Analysis | 59.5% | – |
| LiveBench Language | 78.0% | – |
| LiveBench Instruction Following | 56.8% | – |
| AA Intelligence Index v4.3.2 | 56 | 20.2 |
| IFBench | – | 71.4% |
| AA-LCR | 82.7% | 74.7% |
| MMMU-Pro | – | 70.1% |
| AA-Omniscience | 32.3 | -15.6 |
| Terminal-Bench Hard | – | 37.1% |
| GPQA Diamond (AA) | – | 82.7% |
| Humanity's Last Exam (AA) | 55.0% | 20.1% |
| SciCode (AA) | 61.0% | – |
| 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 | 31.4% | 1.1% |
| GDPval-AA v2.1 | 67.0% | – |
| Analyst Agent (AA) | 57.5% | – |
| FORTRESS | – | 16.0% |
| PropensityBench | – | 10.5% |
| SciPredict | – | 17.9% |
| LMArena Maths | – | 1448 |
| LMArena Creative Writing | – | 1383 |
| LMArena Instruction Following | – | 1403 |
| LMArena Multi-turn | – | 1420 |
| LMArena Longer Queries | – | 1411 |
| Mystery Game Puzzles | 65.0% | – |
| METR Time Horizons | – | 63.6% |
| ForecastBench | – | 62.5% |
| CursorBench | 55.5% | – |
| FrontierSWE | 61.9% | – |
| Terminal-Bench 4.0 (AA) | 63.6% | – |
| AutomationBench | 71.8% | – |
| GDP.pdf | 25.8% | – |
| MLCR | 75.0% | – |
| Harvey LAB | 2.8% | – |
| AA-Omniscience: accuracy | 54.0% | 38.6% |
| AA-Omniscience: non-hallucination | 53.0% | 11.9% |
| AA-Briefcase v1.1 | 1823 | – |
| BrowseComp-Plus | – | 50.5% |
Data as of 2026-10-11. Best configuration of each model; every score links to its source on the model pages.
Claude Sonnet 5.5 vs o3: questions
- Is Claude Sonnet 5.5 better than o3?
- Claude Sonnet 5.5 (max) leads on quality: 67.6 vs 59.7. The BenchLeader Index combines every independent quality benchmark; Claude Sonnet 5.5 (max) is ahead overall as of 2026-10-11, but check the category scores for your use.
- Is Claude Sonnet 5.5 better than o3 for coding?
- Claude Sonnet 5.5 scores higher in coding (70 vs 62 on the category index, where 50 is average).
- Is Claude Sonnet 5.5 better than o3 for agentic tasks?
- o3 scores higher in agentic tasks (70 vs 70 on the category index, where 50 is average).
- Which is cheaper, Claude Sonnet 5.5 or o3?
- o3 is cheaper: $3.50 against $4.00 per million tokens, blended at three input tokens per output token.
- Which is faster, Claude Sonnet 5.5 or o3?
- Claude Sonnet 5.5 streams faster: 141 against 116 output tokens per second.
- Which has the larger context window?
- Claude Sonnet 5.5 accepts more context: 1M against 200k tokens.