gpt-5.5-2026-04-23
OpenAI- average±1.6 (95%)
- 87.0
- leaderboard rankof 114
- ≈8
- benchmarksof 12
- 12
- per 1,000 documentsmean over its benchmarks
- $94
- seconds per inputmean response time
- 40.1
By task
-
Data correction
78.6
Best of 114 models with all its benchmarks: claude-opus-5-5, 82.7.
-
Document understanding
95.5
Best of 107 models with all its benchmarks: gpt-6-astra, 97.0.
-
Information extraction
87.3
Best of 107 models with all its benchmarks: grok-4.6, 90.8.
-
Transcription
91.9
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 | 98.1±1.1 | ≈8of 115 | $109$0.55 run | 35.57 | T1037 · 27 Apr 2026prompt.txt |
| Blacklist Cards | 89.5±4.0 | ≈51of 114 | $29$0.97 run | 6.54 | T1044 · 16 Sep 2026prompt.txt |
| Book Advert XML Repair | 60.0±9.8 | ≈1of 119 | $41$2.04 run | 22.54 | T1047 · 27 Apr 2026prompt.txt |
| Business Letters | 65.0±7.1 | ≈21of 114 | $48$2.74 run | 8.32 | T1038 · 16 Sep 2026prompt.txt |
| Company Lists | 86.3±8.5 | ≈29of 114 | $69$1.03 run | 17.23 | T1045 · 16 Sep 2026prompt.txt |
| Duty Rosters | 92.7±2.5 | ≈5of 107 | $171$3.59 run | 100.59 | T1333 · 24 Jul 2026prompt.txt |
| Fraktur Adverts | 97.1±1.0 | ≈15of 114 | $207$1.03 run | 118.29 | T1041 · 27 Apr 2026prompt_optimized.txt |
| General Meeting Minutes | 90.7±7.1 | ≈1of 108 | $89$0.80 run | 34.79 | T1253 · 3 Jul 2026prompt.txt |
| Library Cards | 88.9±1.3 | ≈15of 114 | $24$6.21 run | 9.73 | T1042 · 27 Apr 2026prompt.txt |
| Magazine Pages | 96.6±2.8 | ≈1of 115 | $51$2.36 run | 15.11 | T1049 · 16 Sep 2026prompt.txt |
| Medieval Manuscripts | 81.6±7.3 | ≈27of 114 | $222$2.67 run | 90.66 | T1043 · 16 Sep 2026prompt.txt |
| Personnel Cards | 97.1±1.2 | ≈30of 114 | $68$4.15 run | 21.86 | T1048 · 27 Apr 2026prompt.txt |
Price
| From | Input, $ per million tokens | Output, $ per million tokens | Source |
|---|---|---|---|
| 27 Apr 2026 | 5.0 | 30.0 | archived price page |
Every run
- T1333Duty Rosters16 Sep 202691.3$3.67
- T1253General Meeting Minutes16 Sep 202689.3$0.84
- T1049Magazine Pages16 Sep 202696.6$2.36
- T1048Personnel Cards16 Sep 202697.0$4.12
- T1047Book Advert XML Repair16 Sep 202651.1$2.12
- T1046Company Lists16 Sep 202674.3$0.88
- T1045Company Lists16 Sep 202686.3$1.03
- T1044Blacklist Cards16 Sep 202689.5$0.97
- T1043Medieval Manuscripts16 Sep 202681.6$2.67
- T1042Library Cards16 Sep 202688.3$6.45
- T1041Fraktur Adverts16 Sep 202697.1$0.83
- T1040Business Letters16 Sep 2026 · 32 signed letters63.0$1.80
- T1039Business Letters16 Sep 2026 · 25 unsigned letters69.0$1.00
- T1038Business Letters16 Sep 2026 · all 57 letters65.0$2.74
- T1037Bibliographic Data16 Sep 202696.1$0.54
- T1333Duty Rosters24 Jul 202692.7$3.59
- T1253General Meeting Minutes3 Jul 202690.7$0.80
- T1049Magazine Pages27 Apr 202695.6$2.37
- T1048Personnel Cards27 Apr 202697.1$4.15
- T1047Book Advert XML Repair27 Apr 202660.0$2.04
- T1046Company Lists27 Apr 202675.5$0.89
- T1045Company Lists27 Apr 202683.6$0.98
- T1044Blacklist Cards27 Apr 202687.8$0.91
- T1043Medieval Manuscripts27 Apr 202673.9$2.33
- T1042Library Cards27 Apr 202688.9$6.21
- T1041Fraktur Adverts27 Apr 202697.1$1.03
- T1040Business Letters27 Apr 2026 · 32 signed letters64.0$1.67
- T1039Business Letters27 Apr 2026 · 25 unsigned letters61.0$1.00
- T1038Business Letters27 Apr 2026 · all 57 letters62.0$2.65
- T1037Bibliographic Data27 Apr 202698.1$0.55