Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Question Answering on LiveMath
Loading...
64.5
Accuracy
SOFTSKILL
-1.748
15.451
32.65
49.849
Jun 18, 2026
Accuracy
Updated 1mo ago
Evaluation Results
Method
Method
Links
Accuracy
SOFTSKILL
Base Model=Qwen3.5–4B,...
2026.06
64.5
SOFTSKILL-mean
Base Model=Qwen3.5–4B,...
2026.06
59.7
LoRA
Base Model=Qwen3.5–4B
2026.06
58.9
SkillOpt
Base Model=Qwen3.5–4B
2026.06
52
SOFTSKILL
Base Model=Qwen3.5–4B,...
2026.06
50
SOFTSKILL-mean
Base Model=Qwen3.5–4B,...
2026.06
42.7
LLM skill
Base Model=Qwen3.5–4B
2026.06
28.8
GEPA
Base Model=Qwen3.5–4B
2026.06
28.8
Trace2Skill
Base Model=Qwen3.5–4B
2026.06
27.2
No skill
Base Model=Qwen3.5–4B
2026.06
22.4
Human skill
Base Model=Qwen3.5–4B
2026.06
18.4
TextGrad
Base Model=Qwen3.5–4B
2026.06
10.6
SOFTSKILL-mean-init-only
Base Model=Qwen3.5–4B,...
2026.06
0.8
Feedback
Search any
task
Search any
task