List your company

Composing auto-generated skills on WebArena

57.6Skill Step PercentageGPT-5
36.90442.27747.6553.023Jun 18, 2026

Evaluation results

MethodLinks
GPT-5Evaluation Protocol=zero-shot format-constrained prompting2026.0657.6
Claude Sonnet 4.5Evaluation Protocol=zero-shot format-constrained prompting2026.0657.6
Llama-3.1-70BEvaluation Protocol=zero-shot format-constrained prompting2026.0656.2
Qwen3-8BEvaluation Protocol=zero-shot2026.0655.8
OLMo-3-7BEvaluation Protocol=zero-shot format-constrained prompting2026.0654.5
Claude Haiku 4.5Evaluation Protocol=zero-shot format-constrained prompting2026.0650.8
DeepSeek-R1-Distill-32BEvaluation Protocol=zero-shot format-constrained prompting2026.0645.1
Qwen3-8BEvaluation Protocol=GRPO from base2026.0644.2
Gemma-4-31BEvaluation Protocol=zero-shot format-constrained prompting2026.0637.7