Benchmark Almanac

Qwen2.5-3B

Qwen2.5-3B.

Qwen Released September 19th, 2024 Family Qwen2.5 The lab's model card Compare with others
22
Benchmarks with a score
0
Ranked first, out of five or more models
0
In the top three
Qwen Sep 2024

Qwen2.5-3B

Open weights Small

No Epoch Capabilities Index score yet
  • Open Arabic LLM Leaderboard v1 43.6% #28 /55
  • AlGhafa 61.1% #18 /39
  • ALRAGE 71.7% #20 /39

Where it ranks

Its best score on each benchmark, and where that ranks among every model with a score there. We rank each model once, at its best setting.

Knowledge and language

BenchmarkScoreRankSource
MuSR 43% #33 of 98 Official leaderboard
AlGhafa 61.1% #18 of 39 Official leaderboard
MadinahQA 47% #18 of 39 Official leaderboard
Open Arabic LLM Leaderboard v2 54.2% #18 of 39 Official leaderboard
Open Arabic LLM Leaderboard v1 43.6% #28 of 55 Official leaderboard
ALRAGE 71.7% #20 of 39 Official leaderboard
ArabicMMLU 48.2% #21 of 39 Official leaderboard
ArbMMLU-HT 35.9% #25 of 39 Official leaderboard
MMLU-CF 55.9% #26 of 37 Official leaderboard
MLogiQA 39.4% #12 of 17 Official leaderboard
MMLU-Pro 43.7% #150 of 212 Official leaderboard
Vectara Hallucination Leaderboard (HHEM) 7.00% #113 of 148 Official leaderboard
HELMET 12.3% #13 of 16 Official leaderboard
SuperGPQA 23.3% #49 of 59 Official leaderboard
Chinese SimpleQA 17.3% #23 of 27 Official leaderboard
EQ-Bench v2 49.8 #78 of 85 Official leaderboard
FaithJudge 56% #34 of 37 Official leaderboard

Math and science

BenchmarkScoreRankSource
ZebraLogic 4.80% #44 of 46 Official leaderboard

Agents and tools

BenchmarkScoreRankSource
C3-Bench 4.30% #23 of 30 Official leaderboard

Vision, audio and robotics

BenchmarkScoreRankSource
Ego3D-Bench 12.6 #10 of 11 Official leaderboard

Safety and security

BenchmarkScoreRankSource
AraTrust 72.5% #22 of 39 Official leaderboard

Other

BenchmarkScoreRankSource
Open LLM Leaderboard v2 18.1 #60 of 99 Official leaderboard