Claude Opus 4.5 vs GPT-5.6 Terra
Claude Opus 4.5 
GPT-5.6 Terra Verdict
- GPT-5.6 Terra (xhigh) leads on quality: 64.4 vs 60.7.
- Claude Opus 4.5 (thinking) is stronger in agents & tools.
- GPT-5.6 Terra (xhigh) is stronger in coding, composite, instruction following, long context, maths, multimodal, human preference.
- GPT-5.6 Terra (xhigh) is 2.2× cheaper ($4.50 vs $10.00 per 1M blended).
- GPT-5.6 Terra (xhigh) streams 1.8× faster (79 vs 45 tokens per second).
| Metric | Claude Opus 4.5 (thinking) | GPT-5.6 Terra (xhigh) |
|---|---|---|
| BenchLeader Index | 60.7 | 64.4 |
| Agents & tools score | 74.6 | 69.8 |
| Coding score | 56.4 | 63.7 |
| Composite score | 65.8 | 77.3 |
| Instruction following score | 54.8 | 65.3 |
| Knowledge score | 61.0 | 61.0 |
| Long context score | 65.2 | 66.0 |
| Maths score | 64.0 | 71.0 |
| Multimodal score | 58.0 | 63.0 |
| Reasoning score | 58.6 | 58.6 |
| Human preference score | – | 66.7 |
| Blended price $/M | $10.00 | $4.50 |
| Output speed | 45 tok/s | 79 tok/s |
| Time to first answer | 17.3 s | 33.6 s |
| Context window | 200k | 1M |
| SimpleBench | – | 48.9% |
| SciCode | – | 51.6% |
| ProofBench | – | 74.0% |
| LMArena Text | – | 1466 |
| LMArena Hard Prompts | – | 1491 |
| LMArena Coding | – | 1519 |
| LMArena WebDev | – | 1521 |
| LMArena Vision | – | 1269 |
| LMArena Agent | – | 1.5 |
| AA Intelligence Index | 29.1 | 38.2 |
| IFBench | 58.0% | 66.3% |
| AA-LCR | 77.3% | 79.0% |
| MMMU-Pro | 74.0% | 79.5% |
| AA-Omniscience | 14 | -3.0 |
| Terminal-Bench Hard | 47.0% | 62.9% |
| GPQA Diamond (AA) | 86.6% | 90.8% |
| Humanity's Last Exam (AA) | 30.1% | 41.9% |
| SciCode (AA) | – | 52.3% |
| τ²-Bench Telecom (AA) | 89.5% | 80.4% |
| AIME (Vals) | 95.4% | – |
| LiveCodeBench | 83.7% | 85.9% |
| MMLU-Pro | 87.3% | 86.7% |
| IOI | 20.3% | 65.3% |
| LegalBench | 84.6% | – |
| CorpFin | 65.1% | 65.3% |
| TaxEval | 74.9% | 76.2% |
| MedQA | 95.9% | – |
| MGSM | 95.2% | – |
| SWE-bench (Vals) | 76.4% | – |
| GPQA Diamond (Vals) | 85.9% | 90.9% |
| MultiChallenge | 59.0% | – |
| PRBench Finance | 46.2% | – |
| PRBench Legal | 44.2% | – |
| VISTA | 46.4% | – |
| MultiNRC | 48.6% | – |
| TutorBench | 51.2% | – |
| Kagi LLM Benchmark | 80.2% | – |
| ARC-AGI-1 | 80.0% | 94.0% |
| ARC-AGI-2 | 37.6% | 74.2% |
| ARC-AGI-3 | – | 0.7% |
Data as of 2026-09-10. Best configuration of each model; every score links to its source on the model pages.
Claude Opus 4.5 vs GPT-5.6 Terra: questions
- Is Claude Opus 4.5 better than GPT-5.6 Terra?
- GPT-5.6 Terra (xhigh) leads on quality: 64.4 vs 60.7. The BenchLeader Index combines every independent quality benchmark; GPT-5.6 Terra (xhigh) is ahead overall as of 2026-09-10, but check the category scores for your use.
- Is Claude Opus 4.5 better than GPT-5.6 Terra for coding?
- GPT-5.6 Terra scores higher in coding (64 vs 56 on the category index, where 50 is average).
- Is Claude Opus 4.5 better than GPT-5.6 Terra for agentic tasks?
- Claude Opus 4.5 scores higher in agentic tasks (75 vs 70 on the category index, where 50 is average).
- Which is cheaper, Claude Opus 4.5 or GPT-5.6 Terra?
- GPT-5.6 Terra is cheaper: $4.50 against $10.00 per million tokens, blended at three input tokens per output token.
- Which is faster, Claude Opus 4.5 or GPT-5.6 Terra?
- GPT-5.6 Terra streams faster: 79 against 45 output tokens per second.
- Which has the larger context window?
- GPT-5.6 Terra accepts more context: 1M against 200k tokens.