FuseChat-Qwen-2.5-7B-Instruct — benchmark results

FuseAI's FuseChat-3.0 knowledge-fusion tune of Qwen2.5-7B-Instruct, distilling strengths of larger source LLMs via SFT and DPO. Provider: Other. Released 2024-11-12. Access: Open.

Unified ELO 1526 ± 69, rank #701 of 1776 rated models, from 7 benchmark results.

Strongest benchmark results

BenchmarkScoreMetricPercentile
Open LLM Leaderboard - MATH Level 545.62Score93.8
AlpacaEval 2.063.58LC Win Rate (%)93.7
Open LLM Leaderboard - MMLU-Pro34.65Score74.9
Open LLM Leaderboard - BBH36.25Score74.2
Open LLM Leaderboard - IFEval59.06Score70.8
Open LLM Leaderboard - GPQA6.15Score51.6
Open LLM Leaderboard - MuSR6.72Score31.9

Interactive version: theaggregate.ai/model?slug=fusechat-qwen-2-5-7b-instruct · How the rankings work · Data refreshed daily, snapshot 2026-07-22.