GPT-5.6 Terra vs o3
GPT-5.6 Terra Verdict
- GPT-5.6 Terra (xhigh) leads on quality: 64.1 vs 61.6.
- GPT-5.6 Terra (xhigh) is stronger in composite, human preference, instruction following, knowledge, long context, multimodal.
- o3 is stronger in agents & tools, coding, maths, reasoning.
- o3 is 1.3× cheaper ($3.50 vs $4.50 per 1M blended).
- o3 streams 1.4× faster (107 vs 76 tokens per second).
| Metric | GPT-5.6 Terra (xhigh) | o3 |
|---|---|---|
| BenchLeader Index | 64.1 | 61.6 |
| Agents & tools score | 69.5 | 70.3 |
| Coding score | 61.5 | 62.0 |
| Composite score | 77.4 | 54.5 |
| Human preference score | 66.5 | 62.3 |
| Instruction following score | 65.3 | 65.0 |
| Knowledge score | 61.0 | 58.0 |
| Long context score | 66.0 | 63.8 |
| Maths score | 71.0 | 78.6 |
| Multimodal score | 63.0 | 57.9 |
| Reasoning score | 58.6 | 61.5 |
| Blended price $/M | $4.50 | $3.50 |
| Output speed | 76 tok/s | 107 tok/s |
| Time to first answer | 30.9 s | 6.3 s |
| Context window | 1.1M | 200k |
| SimpleBench | 48.9% | – |
| SciCode | 51.6% | – |
| ProofBench | 74.0% | – |
| Epoch Capabilities Index | – | 146.9 |
| LMArena Text | 1466 | 1432 |
| LMArena Hard Prompts | 1489 | 1441 |
| LMArena Coding | 1518 | 1460 |
| LMArena WebDev | 1521 | – |
| LMArena Vision | 1269 | 1214 |
| LMArena Agent | 1.5 | – |
| AA Intelligence Index | 38.2 | 20.2 |
| IFBench | 66.3% | 71.4% |
| AA-LCR | 79.0% | 74.7% |
| MMMU-Pro | 79.5% | 70.1% |
| AA-Omniscience | -3.0 | -15.6 |
| Terminal-Bench Hard | 62.9% | 37.1% |
| GPQA Diamond (AA) | 90.8% | 82.7% |
| Humanity's Last Exam (AA) | 41.9% | 20.1% |
| SciCode (AA) | 52.3% | – |
| τ²-Bench Telecom (AA) | 80.4% | 80.7% |
| LiveCodeBench | 85.9% | – |
| MMLU-Pro | 86.7% | – |
| CorpFin | 65.3% | – |
| TaxEval | 76.2% | – |
| GPQA Diamond (Vals) | 90.9% | – |
| 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% |
| ARC-AGI-1 | 94.0% | – |
| ARC-AGI-2 | 74.2% | – |
| ARC-AGI-3 | 0.7% | – |
| BFCL Overall | – | 63.0% |
Data as of 2026-09-12. Best configuration of each model; every score links to its source on the model pages.
GPT-5.6 Terra vs o3: questions
- Is GPT-5.6 Terra better than o3?
- GPT-5.6 Terra (xhigh) leads on quality: 64.1 vs 61.6. The BenchLeader Index combines every independent quality benchmark; GPT-5.6 Terra (xhigh) is ahead overall as of 2026-09-12, but check the category scores for your use.
- Is GPT-5.6 Terra better than o3 for coding?
- o3 scores higher in coding (62 vs 62 on the category index, where 50 is average).
- Is GPT-5.6 Terra 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, GPT-5.6 Terra or o3?
- o3 is cheaper: $3.50 against $4.50 per million tokens, blended at three input tokens per output token.
- Which is faster, GPT-5.6 Terra or o3?
- o3 streams faster: 107 against 76 output tokens per second.
- Which has the larger context window?
- GPT-5.6 Terra accepts more context: 1.1M against 200k tokens.