LIBRA - ruBABILongQA1: leaderboard
LIBRA ruBABILong QA1 subtask: synthetic long-context QA requiring precise retrieval and state tracking over distractor text.
Metric: Dataset Total Score (%). Source: huggingface.co. Saturation forecast: Around January 2027. 13 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Qwen 3 30B A3B 2507 Instruct | 80.5 | #488 |
| 2 | Qwen 3.5 35B A3B | 68.38 | #256 |
| 3 | Qwen 3.5 122B A10B | 66.8 | #173 |
| 4 | Qwen 3.5 9B | 64.88 | #379 |
| 5 | Qwen 3.5 4B | 61.13 | #475 |
| 6 | Llama 3.1 8B Instruct | 60.8 | #1035 |
| 7 | Qwen 3.5 2B | 58.48 | #1148 |
| 8 | Qwen 3.5 0.8B | 55.57 | #1424 |
| 9 | Phi-3.5-mini-instruct | 43.02 | #1167 |
| 10 | LFM2.5-1.2B Instruct | 6.95 | #1462 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=libra-rubabilongqa1 · How It Works · Data refreshed daily, snapshot 2026-10-20.