LIBRA - LongContextMultiQ * — leaderboard

Metric: Dataset Total Score (%). Source: huggingface.co. 13 models tracked.

Top models

#ModelScore
1Qwen 3 30B A3B 2507 Instruct45.96
2Llama 3.1 8B Instruct40.68
3Qwen 3.5 35B A3B39.04
4Qwen 3.5 122B A10B38.14
5Qwen 3.5 9B37.04
6Phi-3.5-mini-instruct34.92
7Qwen 3.5 4B34.9
8Qwen 3.5 2B33.39
9Qwen 3.5 0.8B31.73
10LFM2.5-1.2B Instruct21.2

Interactive version: theaggregate.ai/benchmark?slug=libra-longcontextmultiq-star · How the rankings work · Data refreshed daily, snapshot 2026-07-22.