gemini-3.5-flash-lite
Google- average±2.3 (95%)
- 72.8
- leaderboard rankof 114
- ≈50
- benchmarksof 12
- 12
- per 1,000 documentsmean over its benchmarks
- $2.87
- seconds per inputmean response time
- 12.9
By task
-
Data correction
58.1
Best of 114 models with all its benchmarks: claude-opus-5-5, 82.7.
-
Document understanding
58.8
Best of 107 models with all its benchmarks: gpt-6-astra, 97.0.
-
Information extraction
84.5
Best of 107 models with all its benchmarks: grok-4.6, 90.8.
-
Transcription
86.4
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 | 93.2±3.0 | ≈57of 115 | $6.37$0.032 run | 6.44 | T1435 · 16 Sep 2026prompt.txt |
| Blacklist Cards | 95.3±2.7 | ≈14of 114 | $0.82$0.027 run | 43.76 | T1442 · 25 Jul 2026prompt.txt |
| Book Advert XML Repair | 22.4±7.5 | ≈64of 119 | $1.48$0.074 run | 3.18 | T1445 · 25 Jul 2026prompt.txt |
| Business Letters | 63.0±7.0 | ≈24of 114 | $0.88$0.050 run | 1.78 | T1436 · 16 Sep 2026prompt.txt |
| Company Lists | 89.0±8.8 | ≈6of 114 | $3.20$0.048 run | 3.74 | T1443 · 16 Sep 2026prompt.txt |
| Duty Rosters | 82.5±2.7 | ≈30of 107 | $7.15$0.15 run | 13.87 | T1449 · 25 Jul 2026prompt.txt |
| Fraktur Adverts | 88.6±13.3 | ≈32of 114 | $5.66$0.028 run | 7.02 | T1439 · 16 Sep 2026prompt_optimized.txt |
| General Meeting Minutes | 81.8±11.8 | ≈25of 108 | $2.84$0.026 run | 5.31 | T1448 · 25 Jul 2026prompt.txt |
| Library Cards | 86.9±1.4 | ≈41of 114 | $0.93$0.25 run | 1.29 | T1440 · 16 Sep 2026prompt.txt |
| Magazine Pages | 0.0±0.0 | ≈79of 115 | $0.60$0.028 run | 13.63 | T1447 · 25 Jul 2026prompt.txt |
| Medieval Manuscripts | 76.7±14.3 | ≈27of 114 | $1.08$0.013 run | 2.5 | T1441 · 16 Sep 2026prompt.txt |
| Personnel Cards | 93.9±3.0 | ≈58of 114 | $3.40$0.21 run | 52.52 | T1446 · 25 Jul 2026prompt.txt |
Price
| From | Input, $ per million tokens | Output, $ per million tokens | Source |
|---|---|---|---|
| 25 Jul 2026 | 0.3 | 2.5 | archived price page |
Every run
- T1449Duty Rosters16 Sep 202681.8$0.15
- T1448General Meeting Minutes16 Sep 202680.5$0.026
- T1447Magazine Pages16 Sep 20260.0$0.027
- T1446Personnel Cards16 Sep 202693.7$0.21
- T1445Book Advert XML Repair16 Sep 202618.7$0.075
- T1444Company Lists16 Sep 202672.8$0.045
- T1443Company Lists16 Sep 202689.0$0.048
- T1442Blacklist Cards16 Sep 202694.6$0.027
- T1441Medieval Manuscripts16 Sep 202676.7$0.013
- T1440Library Cards16 Sep 202686.9$0.25
- T1439Fraktur Adverts16 Sep 202688.6$0.028
- T1438Business Letters16 Sep 2026 · 32 signed letters65.0$0.031
- T1437Business Letters16 Sep 2026 · 25 unsigned letters58.0$0.019
- T1436Business Letters16 Sep 2026 · all 57 letters63.0$0.050
- T1435Bibliographic Data16 Sep 202693.2$0.032
- T1449Duty Rosters25 Jul 202682.5$0.15
- T1448General Meeting Minutes25 Jul 202681.8$0.026
- T1447Magazine Pages25 Jul 20260.0$0.028
- T1446Personnel Cards25 Jul 202693.9$0.21
- T1445Book Advert XML Repair25 Jul 202622.4$0.074
- T1444Company Lists25 Jul 202673.9$0.043
- T1443Company Lists25 Jul 202681.3$0.050
- T1442Blacklist Cards25 Jul 202695.3$0.027
- T1441Medieval Manuscripts25 Jul 202674.0$0.013
- T1440Library Cards25 Jul 202686.5$0.25
- T1439Fraktur Adverts25 Jul 202687.4$0.025
- T1438Business Letters25 Jul 2026 · 32 signed letters63.0$0.031
- T1437Business Letters25 Jul 2026 · 25 unsigned letters59.0$0.019
- T1436Business Letters25 Jul 2026 · all 57 letters59.0$0.050
- T1435Bibliographic Data25 Jul 202691.2$0.028