flan-t5-large: benchmark results
Provider: Google. Released 2022-10-21. Access: API.
Unified ELO 1368 ± 1, rank #1224 of 1392 rated models, from 13 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| InstructEval - Problem Solving | 41 | Average (%, MMLU/BBH/DROP/CRASS/HumanEval) | 75.5 |
| InstructEval - Alignment (HHH) | 70.4 | Average (%, harmless/helpful/honest) | 68 |
| ConvRe - Re2Text Hard | 26.2 | Accuracy (%) | 44.4 |
| ConvRe - Text2Re Hard | 29.6 | Accuracy (%) | 44.4 |
| Open LLM Leaderboard - MuSR | 9.01 | Score | 43.5 |
| ConvRe | 51.2 | Average Score (%) | 33.3 |
| ConvRe - Text2Re Easy | 77.3 | Accuracy (%) | 33.3 |
| Open LLM Leaderboard - BBH | 17.51 | Score | 26.4 |
| ConvRe - Re2Text Easy | 71.5 | Accuracy (%) | 22.2 |
| Open LLM Leaderboard - MMLU-Pro | 7.88 | Score | 18.3 |
| Open LLM Leaderboard - IFEval | 22.01 | Score | 15 |
| Open LLM Leaderboard - MATH Level 5 | 1.44 | Score | 10.6 |
Interactive version: theaggregate.ai/model?slug=flan-t5-large · How It Works · Data refreshed daily, snapshot 2026-09-05.