GPT-6.1 Terra (Non-reasoning): benchmark results
Provider: OpenAI. Released 2026-10-26. Access: API.
Aggregate Rating 1846 ± 15, rank #81 of 2178 rated models, from 76 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| AA GDP.pdf - Construction - Criterion Pass Rate | 77.81 | Criterion Pass Rate (%) | 95.1 |
| AA GDP.pdf - Engineering - Criterion Pass Rate | 79.31 | Criterion Pass Rate (%) | 91.4 |
| Wolfram LLM Benchmarking Project | 63 | Correct Functionality (%) | 86.2 |
| AA GDP.pdf - Real Estate - Criterion Pass Rate | 73.01 | Criterion Pass Rate (%) | 85.7 |
| AA FrontierCode 2.0 | 33.35 | Guardrail-gated Rubric Score (%) | 79.6 |
| AA Terminal-Bench Science 0.1 - Earth Sciences | 33.33 | Pass@1 (%) | 79.4 |
| AA GDP.pdf | 23.2 | All Criteria Pass Rate (%) | 78.2 |
| AA GDP.pdf - Healthcare - Criterion Pass Rate | 84.35 | Criterion Pass Rate (%) | 78 |
| AA Terminal-Bench Science 0.1 - Life Sciences | 26.32 | Pass@1 (%) | 77.4 |
| AA Terminal-Bench 4.0 | 26.77 | Pass@1 (%) | 74.9 |
| AA Terminal-Bench Science 0.1 | 24.29 | Pass@1 (%) | 73.4 |
| AA Terminal-Bench Science 0.1 - Mathematical Sciences | 21.57 | Pass@1 (%) | 72.6 |
Interactive version: theaggregate.ai/model?slug=gpt-6-1-terra-non-reasoning · How It Works · Data refreshed daily, snapshot 2026-11-07.