mamba-2.8B-hf: benchmark results
Provider: State Space Models. Released 2024-03-05. Access: Open.
Aggregate Rating 945 ± 44, rank #1656 of 1656 rated models, from 53 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| BABILong (NIAH, v0, 0k) | 70 | QA1–5 Mean Accuracy (%) | 72.4 |
| RoBench-25 (Recovery Instruction) - No Trigger | 0.28 | Clipped Heidke skill score (0-1; chance-corrected accuracy o | 40 |
| RoBench-25 (Recovery Instruction) - Trigger A | 0.22 | Clipped Heidke skill score (0-1; chance-corrected accuracy o | 40 |
| RoBench-25 (Recovery Instruction) - Trigger B | 0.08 | Clipped Heidke skill score (0-1; chance-corrected accuracy o | 40 |
| RoBench-25 (Recovery Instruction) - Trigger F | 0.15 | Clipped Heidke skill score (0-1; chance-corrected accuracy o | 40 |
| RoBench-25 (Recovery Instruction) - Trigger G | 0.28 | Clipped Heidke skill score (0-1; chance-corrected accuracy o | 40 |
| RoBench-25 (Awareness Instruction) - No Trigger | 0.19 | Clipped Heidke skill score (0-1; chance-corrected accuracy o | 33.3 |
| RoBench-25 (Awareness Instruction) - Trigger A | 0.18 | Clipped Heidke skill score (0-1; chance-corrected accuracy o | 33.3 |
| RoBench-25 (Awareness Instruction) - Trigger B | 0.07 | Clipped Heidke skill score (0-1; chance-corrected accuracy o | 33.3 |
| RoBench-25 (Awareness Instruction) - Trigger C | 0.15 | Clipped Heidke skill score (0-1; chance-corrected accuracy o | 33.3 |
| RoBench-25 (Awareness Instruction) - Trigger D | 0.08 | Clipped Heidke skill score (0-1; chance-corrected accuracy o | 33.3 |
| RoBench-25 (Awareness Instruction) - Trigger E | 0.14 | Clipped Heidke skill score (0-1; chance-corrected accuracy o | 33.3 |
Interactive version: theaggregate.ai/model?slug=mamba-2-8b-hf · How It Works · Data refreshed daily, snapshot 2026-10-20.