OLMo-2-1124-7B-Instruct: benchmark results
Allen AI's instruct variant of the fully open OLMo 2 7B, post-trained with a Tulu 3-style recipe of SFT, DPO, and RLVR. Provider: Allen AI. Released 2024-11-26. Access: Open.
Aggregate Rating 1358 ± 13, rank #1367 of 1656 rated models, from 257 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| LA Leaderboard - PAWS-X Spanish | 66.7 | Accuracy (%) | 95.6 |
| LA Leaderboard - MGSM Spanish | 13.6 | Exact Match (%) | 95.4 |
| Grip on LLMs - Dutch Simplification - Amsterdam | 43.47 | SARI (0-100) | 90 |
| LA Leaderboard - PAWS Catalan | 68.25 | Accuracy (%) | 89.7 |
| LA Leaderboard - PAWS Galician | 65.7 | Accuracy (%) | 89 |
| LA Leaderboard - WNLI Catalan | 64.79 | Accuracy (%) | 89 |
| LA Leaderboard - XCOPA Basque | 53.8 | Accuracy (%) | 87.5 |
| Grip on LLMs - Dutch Simplification - INT Duidelijke Taal | 41.28 | SARI (0-100) | 86.7 |
| Grip on LLMs - Dutch Simplification Average | 42.38 | SARI (0-100; mean of two datasets) | 86.7 |
| Open LLM Leaderboard - IFEval | 72.44 | Score | 86.3 |
| LA Leaderboard - ParafrasesGL | 54.08 | Accuracy (%) | 86 |
| LA Leaderboard - Fake News Spanish | 73.95 | Accuracy (%) | 85.3 |
Interactive version: theaggregate.ai/model?slug=olmo-2-1124-7b-instruct · How It Works · Data refreshed daily, snapshot 2026-10-20.