GPT-5.6 Terra occupies the middle ground of OpenAI's three-tier GPT-5.6 family — strong enough for serious coding, analysis, and content generation but at half the cost of Sol. It shares the same 1M+ context window and 128K output limit. For most production applications, Terra is the right starting point: try it first and only move to Sol if you hit quality ceilings on your hardest prompts. At $2.50/$15 per million tokens, it undercuts GPT-5.5 significantly while matching or exceeding its performance.
| Benchmark | Score | Type | Recorded |
|---|---|---|---|
| GPQA Diamond | 92.9 | accuracy | 29d ago |
| LiveCodeBench | 85.9 | accuracy | 29d ago |
| IFBench | 71.2 | accuracy | 29d ago |
| MMLU-Pro | 86.7 | accuracy | 29d ago |
| LCR | 83.0 | accuracy | 29d ago |
| TAU2 | 86.3 | accuracy | 29d ago |
| SWE-Bench | 95.4 | accuracy | 29d ago |
| SciCode | 55.0 | accuracy | 29d ago |
| Humanity's Last Exam | 42.9 | accuracy | 29d ago |