claude-opus-4-8
Anthropic- average±2.1 (95%)
- 82.2
- leaderboard rankof 114
- ≈14
- benchmarksof 12
- 12
- per 1,000 documentsmean over its benchmarks
- $48
- seconds per inputmean response time
- 13.4
By task
-
Data correction
63.9
Best of 114 models with all its benchmarks: claude-opus-5-5, 82.7.
-
Document understanding
88.5
Best of 107 models with all its benchmarks: gpt-6-astra, 97.0.
-
Information extraction
84.6
Best of 107 models with all its benchmarks: grok-4.6, 90.8.
-
Transcription
94.1
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 | 98.2±1.4 | ≈8of 115 | $78$0.39 run | 19.28 | T1076 · 16 Sep 2026prompt.txt |
| Blacklist Cards | 87.2±3.6 | ≈79of 114 | $28$0.92 run | 3.93 | T1083 · 16 Sep 2026prompt.txt |
| Book Advert XML Repair | 30.4±9.3 | ≈41of 119 | $37$1.86 run | 12.4 | T1086 · 4 Jun 2026prompt.txt |
| Business Letters | 77.0±6.1 | ≈1of 114 | $42$2.41 run | 4.08 | T1077 · 4 Jun 2026prompt.txt |
| Company Lists | 67.8±15.1 | ≈84of 114 | $55$0.82 run | 13.33 | T1084 · 4 Jun 2026prompt.txt |
| Duty Rosters | 86.5±2.5 | ≈18of 107 | $47$0.99 run | 16.84 | T1336 · 17 Sep 2026prompt.txt |
| Fraktur Adverts | 99.2±0.2 | ≈7of 114 | $92$0.46 run | 44.62 | T1080 · 16 Sep 2026prompt_optimized.txt |
| General Meeting Minutes | 86.4±8.7 | ≈25of 108 | $55$0.49 run | 14.55 | T1215 · 3 Jul 2026prompt.txt |
| Library Cards | 88.9±1.4 | ≈15of 114 | $21$5.60 run | 3.87 | T1081 · 16 Sep 2026prompt.txt |
| Magazine Pages | 81.6±9.8 | ≈18of 115 | $26$1.17 run | 3.59 | T1088 · 4 Jun 2026prompt.txt |
| Medieval Manuscripts | 85.8±9.0 | ≈1of 114 | $34$0.41 run | 12.21 | T1082 · 4 Jun 2026prompt.txt |
| Personnel Cards | 97.3±0.9 | ≈1of 114 | $67$4.07 run | 12.37 | T1087 · 16 Sep 2026prompt.txt |
Price
| From | Input, $ per million tokens | Output, $ per million tokens | Source |
|---|---|---|---|
| 4 Jun 2026 | 5.0 | 25.0 | archived price page |
Every run
- T1336Duty Rosters17 Sep 202686.5$0.99
- T1215General Meeting Minutes17 Sep 202686.0$0.50
- T1079Business Letters17 Sep 2026 · 32 signed letters73.0$1.52
- T1078Business Letters17 Sep 2026 · 25 unsigned letters93.0$0.89
- T1077Business Letters17 Sep 2026 · all 57 letters75.0$2.41
- T1088Magazine Pages16 Sep 202672.2$1.17
- T1087Personnel Cards16 Sep 202697.3$4.07
- T1086Book Advert XML Repair16 Sep 202627.3$1.86
- T1085Company Lists16 Sep 202663.3$0.79
- T1084Company Lists16 Sep 202667.6$0.82
- T1083Blacklist Cards16 Sep 202687.2$0.92
- T1082Medieval Manuscripts16 Sep 202683.5$0.41
- T1081Library Cards16 Sep 202688.9$5.60
- T1080Fraktur Adverts16 Sep 202699.2$0.46
- T1079Business Letters16 Sep 2026 · 32 signed letters73.0$1.52
- T1078Business Letters16 Sep 2026 · 25 unsigned letters93.0$0.90
- T1077Business Letters16 Sep 2026 · all 57 letters76.0$2.41
- T1076Bibliographic Data16 Sep 202698.2$0.39
- T1336Duty Rosters24 Jul 202685.6$1.00
- T1215General Meeting Minutes3 Jul 202686.4$0.49
- T1088Magazine Pages4 Jun 202681.6$1.17
- T1087Personnel Cards4 Jun 202692.0$3.99
- T1086Book Advert XML Repair4 Jun 202630.4$1.86
- T1085Company Lists4 Jun 202664.9$0.79
- T1084Company Lists4 Jun 202667.8$0.82
- T1083Blacklist Cards4 Jun 202686.0$0.93
- T1082Medieval Manuscripts4 Jun 202685.8$0.41
- T1081Library Cards4 Jun 202688.6$5.61
- T1080Fraktur Adverts4 Jun 202679.5$0.47
- T1079Business Letters4 Jun 2026 · 32 signed letters73.0$1.51
- T1078Business Letters4 Jun 2026 · 25 unsigned letters93.0$0.90
- T1077Business Letters4 Jun 2026 · all 57 letters77.0$2.41
- T1076Bibliographic Data4 Jun 202697.5$0.38