Tulu 3 405B ranks first on PopQA.
Ai2 Jan 2025 Open weights Flagship
Its best score on each benchmark, and where that ranks among every model with a score there. We rank each model once, at its best setting.
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| BIG-Bench Hard | 88.6% | #9 of 73 | Lab model card | |
| PopQA | 55.5% | #1 of 7 | Lab model card | |
| MMLU | 87% | #41 of 131 | Lab model card | |
| AlpacaEval 2.0 | 51.4% | #8 of 16 | Lab model card | |
| IFEval | 86% | #73 of 135 | Lab model card |
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| HumanEval | 95.9% | #4 of 101 | Lab model card | |
| HumanEval+ | 92.9% | #2 of 27 | Lab model card |
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| Tulu 3 Safety Eval | 86.7% | #2 of 6 | Lab model card |
The Self-Reported Index fits one score to the 8 capability results Ai2 reported for this model, on the scale of the Epoch Capabilities Index. Labs choose which benchmarks to show and how to run them, so read it as the lab's claim, not an independent measurement. The 90% interval runs from 123.0 to 131.2. How we compute it.
The benchmarks the lab chose to headline, with the scores it gave.
Every result in the longer documents the lab published with it.