claude-sonnet-5
Anthropic- average±3.3 (95%)
- 42.9
- leaderboard rankof 114
- 108
- benchmarksof 12
- 12
- per 1,000 documentsmean over its benchmarks
- $19
- seconds per inputmean response time
- 10.4
By task
-
Data correction
46.6
Best of 114 models with all its benchmarks: claude-opus-5-5, 82.7.
-
Document understanding
41.6
Best of 107 models with all its benchmarks: gpt-6-astra, 97.0.
-
Information extraction
59.6
Best of 107 models with all its benchmarks: grok-4.6, 90.8.
-
Transcription
19.3
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 | 98.0±1.6 | ≈8of 115 | $32$0.16 run | 15.74 | T1177 · 26 Sep 2026prompt.txt |
| Blacklist Cards | 39.9±12.0 | ≈110of 114 | $11$0.36 run | 4.06 | T1184 · 1 Jul 2026prompt.txt |
| Book Advert XML Repair | 35.2±8.2 | ≈41of 119 | $15$0.76 run | 10.39 | T1187 · 16 Sep 2026prompt.txt |
| Business Letters | 60.0±6.1 | ≈24of 114 | $17$0.98 run | 3.12 | T1178 · 17 Sep 2026prompt.txt |
| Company Lists | 80.3±11.3 | ≈58of 114 | $21$0.32 run | 10.65 | T1185 · 1 Jul 2026prompt.txt |
| Duty Rosters | 62.1±15.0 | ≈87of 107 | $18$0.37 run | 13.41 | T1343 · 24 Jul 2026prompt.txt |
| Fraktur Adverts | 0.0±0.0 | ≈108of 114 | $33$0.17 run | 32.09 | T1181 · 1 Jul 2026prompt_optimized.txt |
| General Meeting Minutes | 67.7±25.9 | ≈85of 108 | $21$0.19 run | 9.98 | T1218 · 3 Jul 2026prompt.txt |
| Library Cards | 9.2±3.3 | 114of 114 | $8.58$2.26 run | 3.29 | T1182 · 16 Sep 2026prompt.txt |
| Magazine Pages | 4.8±5.1 | ≈79of 115 | $10$0.47 run | 3.49 | T1189 · 1 Jul 2026prompt.txt |
| Medieval Manuscripts | 0.0±0.0 | ≈112of 114 | $14$0.17 run | 9.5 | T1183 · 1 Jul 2026prompt.txt |
| Personnel Cards | 58.0±14.7 | ≈104of 114 | $26$1.60 run | 9.39 | T1188 · 16 Sep 2026prompt.txt |
Price
| From | Input, $ per million tokens | Output, $ per million tokens | Source |
|---|---|---|---|
| 1 Jul 2026 | 2.0 | 10.0 | archived price page |
Every run
- T1177Bibliographic Data26 Sep 202698.0$0.16
- T1343Duty Rosters17 Sep 202630.5$0.38
- T1218General Meeting Minutes17 Sep 202646.9$0.20
- T1180Business Letters17 Sep 2026 · 32 signed letters64.0$0.61
- T1179Business Letters17 Sep 2026 · 25 unsigned letters55.0$0.37
- T1178Business Letters17 Sep 2026 · all 57 letters60.0$0.98
- T1189Magazine Pages16 Sep 20262.2$0.47
- T1188Personnel Cards16 Sep 202658.0$1.60
- T1187Book Advert XML Repair16 Sep 202635.2$0.76
- T1186Company Lists16 Sep 202674.7$0.30
- T1185Company Lists16 Sep 202673.0$0.32
- T1184Blacklist Cards16 Sep 202623.2$0.36
- T1183Medieval Manuscripts16 Sep 20260.0$0.17
- T1182Library Cards16 Sep 20269.2$2.26
- T1181Fraktur Adverts16 Sep 20260.0$0.19
- T1180Business Letters16 Sep 2026 · 32 signed letters63.0$0.61
- T1179Business Letters16 Sep 2026 · 25 unsigned letters55.0$0.37
- T1178Business Letters16 Sep 2026 · all 57 letters59.0$0.98
- T1177Bibliographic Data16 Sep 202639.1$0.12
- T1343Duty Rosters24 Jul 202662.1$0.37
- T1218General Meeting Minutes3 Jul 202667.7$0.19
- T1189Magazine Pages1 Jul 20264.8$0.47
- T1188Personnel Cards1 Jul 202655.3$1.61
- T1187Book Advert XML Repair1 Jul 202628.8$0.78
- T1186Company Lists1 Jul 202677.7$0.30
- T1185Company Lists1 Jul 202680.3$0.32
- T1184Blacklist Cards1 Jul 202639.9$0.36
- T1183Medieval Manuscripts1 Jul 20260.0$0.17
- T1182Library Cards1 Jul 20264.7$2.27
- T1181Fraktur Adverts1 Jul 20260.0$0.17
- T1180Business Letters1 Jul 2026 · 32 signed letters62.0$0.61
- T1179Business Letters1 Jul 2026 · 25 unsigned letters54.0$0.37
- T1178Business Letters1 Jul 2026 · all 57 letters58.0$0.98
- T1177Bibliographic Data1 Jul 202697.7$0.16