ZYao720/WebArbiter-8B-Qwen3
Text Generation • 8B • Updated • 157
WebArbiter - Models
WebArbiter process reward models for web agents. Reasoning distillation + RL. ICLR 2026.
Note Strongest — Avg. BoN Acc 76.66% (Qwen3-8B backbone)
ZYao720/WebArbiter-7B
Text Generation • 8B • Updated • 235Note Flagship paper model — Avg. BoN Acc 74.60%, outperforms GPT-5 by 9.1 pts (Qwen2.5-7B-Instruct)
ZYao720/WebArbiter-4B-Qwen3
Text Generation • 4B • Updated • 157Note Efficient — Avg. BoN Acc 72.55% with ~half the params of 7B (Qwen3-4B)
ZYao720/WebArbiter-3B
Text Generation • 3B • Updated • 154Note Compact — Avg. BoN Acc 59.06%, outperforms WebShepherd-8B (Qwen2.5-3B-Instruct)