claude-haiku-4-5-20251001
Anthropic- average±3.4 (95%)
- 63.9
- leaderboard rankof 114
- ≈79
- benchmarksof 12
- 12
- per 1,000 documentsmean over its benchmarks
- $10
- seconds per inputmean response time
- 11.6
By task
-
Data correction
54.6
Best of 114 models with all its benchmarks: claude-opus-5-5, 82.7.
-
Document understanding
56.8
Best of 107 models with all its benchmarks: gpt-6-astra, 97.0.
-
Information extraction
76.0
Best of 107 models with all its benchmarks: grok-4.6, 90.8.
-
Transcription
71.1
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 | 96.7±2.3 | ≈40of 115 | $13$0.065 run | 14.51 | T0521 · 16 Sep 2026prompt.txt |
| Blacklist Cards | 82.7±3.9 | ≈95of 114 | $3.80$0.13 run | 2.49 | T0528 · 24 Jan 2026prompt.txt |
| Book Advert XML Repair | 19.3±6.0 | ≈64of 119 | $5.39$0.27 run | 5.32 | T0531 · 24 Jan 2026prompt.txt |
| Business Letters | 47.0±7.2 | ≈104of 114 | $4.42$0.25 run | 2.06 | T0522 · 16 Sep 2026prompt.txt |
| Company Lists | 78.3±12.6 | ≈58of 114 | $8.18$0.12 run | 9.46 | T0529 · 16 Sep 2026prompt.txt |
| Duty Rosters | 78.4±3.2 | ≈46of 107 | $15$0.32 run | 12.46 | T1276 · 17 Sep 2026prompt.txt |
| Fraktur Adverts | 58.1±32.2 | ≈67of 114 | $45$0.22 run | 65.39 | T0525 · 16 Sep 2026prompt_optimized.txt |
| General Meeting Minutes | 65.3±13.0 | ≈85of 108 | $7.79$0.070 run | 9.58 | T1210 · 3 Jul 2026prompt.txt |
| Library Cards | 83.8±1.6 | ≈73of 114 | $3.69$0.97 run | 2.18 | T0526 · 24 Jan 2026prompt.txt |
| Magazine Pages | 2.2±4.3 | ≈79of 115 | $2.89$0.13 run | 1.8 | T0806 · 16 Sep 2026prompt.txt |
| Medieval Manuscripts | 65.3±14.0 | ≈72of 114 | $4.52$0.054 run | 6.43 | T0527 · 16 Sep 2026prompt.txt |
| Personnel Cards | 89.9±3.1 | ≈78of 114 | $9.08$0.55 run | 7.57 | T0576 · 16 Sep 2026prompt.txt |
Price
| From | Input, $ per million tokens | Output, $ per million tokens | Source |
|---|---|---|---|
| 23 Jan 2026 | 1 | 5.0 | archived price page |
Every run
- T1276Duty Rosters17 Sep 202678.4$0.32
- T1210General Meeting Minutes17 Sep 202665.0$0.071
- T0524Business Letters17 Sep 2026 · 32 signed letters48.0$0.15
- T0523Business Letters17 Sep 2026 · 25 unsigned letters43.0$0.098
- T0522Business Letters17 Sep 2026 · all 57 letters47.0$0.25
- T0806Magazine Pages16 Sep 20262.2$0.13
- T0576Personnel Cards16 Sep 202689.9$0.55
- T0531Book Advert XML Repair16 Sep 202618.6$0.27
- T0530Company Lists16 Sep 202667.3$0.12
- T0529Company Lists16 Sep 202678.3$0.12
- T0528Blacklist Cards16 Sep 202681.3$0.13
- T0527Medieval Manuscripts16 Sep 202665.3$0.054
- T0526Library Cards16 Sep 202683.7$0.97
- T0525Fraktur Adverts16 Sep 202658.1$0.22
- T0524Business Letters16 Sep 2026 · 32 signed letters49.0$0.15
- T0523Business Letters16 Sep 2026 · 25 unsigned letters43.0$0.098
- T0522Business Letters16 Sep 2026 · all 57 letters47.0$0.25
- T0521Bibliographic Data16 Sep 202696.7$0.065
- T1276Duty Rosters24 Jul 202678.2$0.32
- T1210General Meeting Minutes3 Jul 202665.3$0.070
- T0806Magazine Pages23 Mar 20260.0$0.13
- T0576Personnel Cards9 Feb 202685.5$0.54
- T0531Book Advert XML Repair24 Jan 202619.3$0.27
- T0530Company Lists24 Jan 202663.6$0.12
- T0529Company Lists24 Jan 202677.8$0.13
- T0528Blacklist Cards24 Jan 202682.7$0.13
- T0527Medieval Manuscripts24 Jan 202663.0$0.054
- T0526Library Cards24 Jan 202683.8$0.97
- T0525Fraktur Adverts24 Jan 202611.1$0.040
- T0524Business Letters24 Jan 2026 · 32 signed letters47.0$0.15
- T0523Business Letters24 Jan 2026 · 25 unsigned letters43.0$0.098
- T0522Business Letters24 Jan 2026 · all 57 letters45.0$0.25
- T0521Bibliographic Data24 Jan 202696.2$0.067