OPENAI · MODEL TELEMETRY CARD
GPT-5.6 Terra
Canonical UID: openai/gpt-5.6-terra · Released: 2026-07-09
| Domain | Benchmark | Effort / Mode | ⚡ Value/Cost # | 🏆 Performance # | Score | Task $ | Unit Cost $ | Unit |
|---|---|---|---|---|---|---|---|---|
| Code | ALE-Bench | MAX | #14 / 94 | #9 / 94 | 1951.42 | $1.25405 | $0.0006426 | rating point |
| Reason | ProofBench | XHIGH | #17 / 62 | #15 / 62 | 0.74 | $1.27421 | $1.722 | correct answer |
| Agent | DeepSWE | MEDIUM | #56 / 65 | #57 / 65 | 0.351 | $0.583249 | $1.661 | solved task |
| Agent | DeepSWE | LOW | #57 / 65 | #62 / 65 | 0.241 | $0.427745 | $1.778 | solved task |
| Agent | DeepSWE | HIGH | #2 / 65 | #40 / 65 | 0.538 | $1.13438 | $2.11 | solved task |
| Agent | DeepSWE | XHIGH | #14 / 65 | #30 / 65 | 0.602 | $2.12719 | $3.535 | solved task |
| Agent | DeepSWE | MAX | #33 / 65 | #14 / 65 | 0.696 | $4.94585 | $7.104 | solved task |
Compare against all models on ← #frontier Live Ranking Board · Fixed-capability cost half-life analysis on /methodology.