perceptron/perceptron-mk1.5
Perceptronvia OpenRouter- average±3.3 (95%)
- 57.0
- leaderboard rankof 114
- ≈92
- benchmarksof 12
- 12
- per 1,000 documentsmean over its benchmarks
- $1.63
- seconds per inputmean response time
- 10.6
By task
-
Data correction
54.4
Best of 114 models with all its benchmarks: claude-opus-5-5, 82.7.
-
Document understanding
55.8
Best of 107 models with all its benchmarks: gpt-6-astra, 97.0.
-
Information extraction
72.0
Best of 107 models with all its benchmarks: grok-4.6, 90.8.
-
Transcription
55.9
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 Dataother prompt | 58.9±29.6 | ≈103of 115 | $4.21$0.021 run | 53.82 | T2124 · 24 Oct 2026prompt_v2.txt |
| Blacklist Cards | 91.2±3.5 | ≈51of 114 | $0.63$0.021 run | 4.63 | T2131 · 24 Oct 2026prompt.txt |
| Book Advert XML Repairother prompt | 11.8±5.8 | ≈89of 119 | $1.20$0.060 run | 4.01 | T2133 · 24 Oct 2026prompt_v2.txt |
| Business Letters | 52.0±7.0 | ≈95of 114 | $0.93$0.053 run | 6.33 | T2125 · 24 Oct 2026prompt.txt |
| Company Listsother prompt | 75.5±11.2 | ≈58of 114 | $1.97$0.030 run | 7.76 | T2132 · 24 Oct 2026prompt_v2.txt |
| Duty Rosters | 70.4±3.3 | ≈78of 107 | $2.67$0.056 run | 8.23 | T2137 · 24 Oct 2026prompt.txt |
| Fraktur Adverts | 0.0±0.0 | ≈108of 114 | $2.42$0.012 run | 9.28 | T2128 · 24 Oct 2026prompt_optimized.txt |
| General Meeting Minutes | 75.0±14.5 | ≈85of 108 | $1.54$0.014 run | 13.02 | T2136 · 24 Oct 2026prompt.txt |
| Library Cardsother prompt | 81.1±1.9 | ≈81of 114 | $0.44$0.12 run | 2.09 | T2129 · 24 Oct 2026prompt_year-string.txt |
| Magazine Pagesother prompt | 0.0±0.0 | ≈79of 115 | $0.62$0.028 run | 5.14 | T2135 · 24 Oct 2026prompt_explicit-json.txt |
| Medieval Manuscripts | 70.5±14.8 | ≈27of 114 | $0.83$0.010 run | 4.9 | T2130 · 24 Oct 2026prompt.txt |
| Personnel Cards | 97.1±1.0 | ≈30of 114 | $2.14$0.13 run | 7.86 | T2134 · 24 Oct 2026prompt.txt |
Price
| From | Input, $ per million tokens | Output, $ per million tokens | Source |
|---|---|---|---|
| 24 Oct 2026 | 0.15 | 1.5 | archived price page |
Every run
- T2137Duty Rosters24 Oct 202670.4$0.056
- T2136General Meeting Minutes24 Oct 202675.0$0.014
- T2135Magazine Pages24 Oct 20260.0$0.028
- T2134Personnel Cards24 Oct 202697.1$0.13
- T2133Book Advert XML Repair24 Oct 202611.8$0.060
- T2132Company Lists24 Oct 202675.5$0.030
- T2131Blacklist Cards24 Oct 202691.2$0.021
- T2130Medieval Manuscripts24 Oct 202670.5$0.010
- T2129Library Cards24 Oct 202681.1$0.12
- T2128Fraktur Adverts24 Oct 20260.0$0.012
- T2127Business Letters24 Oct 2026 · 32 signed letters54.0$0.033
- T2126Business Letters24 Oct 2026 · 25 unsigned letters43.0$0.020
- T2125Business Letters24 Oct 2026 · all 57 letters52.0$0.053
- T2124Bibliographic Data24 Oct 202658.9$0.021