FuseChat-Qwen-2.5-7B-Instruct

FuseAI

Parameters

7.6B

Downloads

—

Likes

—

Avg score

31.4

6 benchmarks

Weights

Open

Released

—

About

FuseChat-Qwen-2.5-7B-Instruct is a 7.6B-parameter open-weights model in the FuseChat family from FuseAI. Benchmarked on 6 evals, averaging 31.4.

Benchmark scores · 6

avg 31.4 — higher is better

IFEvalOpen LLM LB · 2026-05

59.1

MATH Lvl 5Open LLM LB · 2026-05

45.6

BIG-Bench HardOpen LLM LB · 2026-05

36.3

MMLU-ProOpen LLM LB · 2026-05

34.6

MuSROpen LLM LB · 2026-05

6.7

GPQA DiamondOpen LLM LB · 2026-05

6.2

Details

PublisherFuseAI

FamilyFuseChat

Parameters7.6B

License

WeightsOpen

Released—

History

Capability over time

Append-only ledger — every observed change to this model's numbers.