gpt-5.6-terra
OpenAI- average±1.7 (95%)
- 82.5
- leaderboard rankof 114
- ≈14
- benchmarksof 12
- 12
- per 1,000 documentsmean over its benchmarks
- $24
- seconds per inputmean response time
- 21.5
By task
-
Data correction
71.0
Best of 114 models with all its benchmarks: claude-opus-5-5, 82.7.
-
Document understanding
92.8
Best of 107 models with all its benchmarks: gpt-6-astra, 97.0.
-
Information extraction
84.2
Best of 107 models with all its benchmarks: grok-4.6, 90.8.
-
Transcription
86.9
Best of 114 models with all its benchmarks: gemini-4-argon, 96.2.
By benchmark
| Benchmark | ScoreThe best current run on each benchmark, on the benchmark's metric scaled to 0–100. ± is the 95% sampling margin, from resampling the run's inputs. | RankRank among all current models on that benchmark; ≈ for models that cannot be told apart within sampling error. | Per 1,000 docs | s / input | Best run |
|---|---|---|---|---|---|
| Bibliographic Data | 97.1±1.9 | ≈29of 115 | $36$0.18 run | 55.35 | T1360 · 16 Sep 2026prompt.txt |
| Blacklist Cards | 91.7±3.4 | ≈51of 114 | $9.80$0.32 run | 4.9 | T1367 · 16 Sep 2026prompt.txt |
| Book Advert XML Repair | 45.2±9.5 | ≈16of 119 | $10$0.52 run | 9.74 | T1370 · 16 Sep 2026prompt.txt |
| Business Letters | 56.0±6.4 | ≈75of 114 | $15$0.83 run | 6.85 | T1361 · 16 Sep 2026prompt.txt |
| Company Lists | 85.5±8.4 | ≈29of 114 | $24$0.36 run | 12.82 | T1368 · 25 Jul 2026prompt.txt |
| Duty Rosters | 86.8±3.2 | ≈18of 107 | $44$0.92 run | 41.88 | T1374 · 16 Sep 2026prompt.txt |
| Fraktur Adverts | 94.4±1.6 | ≈32of 114 | $50$0.25 run | 61.27 | T1364 · 16 Sep 2026prompt_optimized.txt |
| General Meeting Minutes | 87.2±7.5 | ≈25of 108 | $33$0.30 run | 16.59 | T1373 · 25 Jul 2026prompt.txt |
| Library Cards | 84.8±1.5 | ≈51of 114 | $8.50$2.24 run | 4.16 | T1365 · 25 Jul 2026prompt.txt |
| Magazine Pages | 95.0±5.0 | ≈1of 115 | $10$0.47 run | 12.14 | T1372 · 16 Sep 2026prompt.txt |
| Medieval Manuscripts | 69.5±10.6 | ≈27of 114 | $24$0.28 run | 23.34 | T1366 · 16 Sep 2026prompt.txt |
| Personnel Cards | 96.8±1.2 | ≈30of 114 | $27$1.65 run | 9.48 | T1371 · 25 Jul 2026prompt.txt |
Price
| From | Input, $ per million tokens | Output, $ per million tokens | Source |
|---|---|---|---|
| 16 Sep 2026 | 2.0 | 12.0 | archived price page |
| 25 Jul 2026 | 2.5 | 15 | archived price page |
Every run
- T1374Duty Rosters16 Sep 202686.8$0.92
- T1373General Meeting Minutes16 Sep 202686.8$0.23
- T1372Magazine Pages16 Sep 202695.0$0.47
- T1371Personnel Cards16 Sep 202696.8$1.37
- T1370Book Advert XML Repair16 Sep 202645.2$0.52
- T1369Company Lists16 Sep 202672.8$0.26
- T1368Company Lists16 Sep 202685.5$0.30
- T1367Blacklist Cards16 Sep 202691.7$0.32
- T1366Medieval Manuscripts16 Sep 202669.5$0.28
- T1365Library Cards16 Sep 202676.5$2.96
- T1364Fraktur Adverts16 Sep 202694.4$0.25
- T1363Business Letters16 Sep 2026 · 32 signed letters59.0$0.52
- T1362Business Letters16 Sep 2026 · 25 unsigned letters53.0$0.30
- T1361Business Letters16 Sep 2026 · all 57 letters56.0$0.83
- T1360Bibliographic Data16 Sep 202697.1$0.18
- T1374Duty Rosters25 Jul 202686.8$1.15
- T1373General Meeting Minutes25 Jul 202687.2$0.30
- T1372Magazine Pages25 Jul 202694.6$0.56
- T1371Personnel Cards25 Jul 202696.8$1.65
- T1370Book Advert XML Repair25 Jul 202640.0$0.64
- T1369Company Lists25 Jul 202678.1$0.34
- T1368Company Lists25 Jul 202685.5$0.36
- T1367Blacklist Cards25 Jul 202690.6$0.39
- T1366Medieval Manuscripts25 Jul 202667.5$0.50
- T1365Library Cards25 Jul 202684.8$2.24
- T1364Fraktur Adverts25 Jul 202693.0$0.24
- T1363Business Letters25 Jul 2026 · 32 signed letters59.0$0.63
- T1362Business Letters25 Jul 2026 · 25 unsigned letters46.0$0.36
- T1361Business Letters25 Jul 2026 · all 57 letters55.0$0.98
- T1360Bibliographic Data25 Jul 202695.8$0.24