List your company

Multi-turn tool use on GAIA

76.4Pass@1WebAnchor-30B
18.1633.2848.463.52Jan 6, 2026

Evaluation results

MethodLinks
WebAnchor-30BRL Algo=Anchor-GRPO2026.0176.486.4
WebAnchor-30BRL Algo=GRPO2026.0175.887.6
WebAnchor-30BRL Algo=First-step GRPO2026.0174.887.4
OpenAI-o3†RL Algo=-2026.0170.5—
Claude-4-Sonnet†RL Algo=-2026.0168.3—
DeepSeek-V3.1†RL Algo=-2026.0163.1—
Kimi-K2†RL Algo=-2026.0157.7—
WebSailor-32BRL Algo=-2026.0153.2—
WebDancer-QwQRL Algo=-2026.0151.5—
WebThinker-RLRL Algo=-2026.0148.5—
WebAnchor-7BRL Algo=First-step GRPO2026.0138.555.1
WebSailor-7BRL Algo=-2026.0137.9—
WebAnchor-7BRL Algo=Anchor-GRPO2026.0137.851.5
WebAnchor-7BRL Algo=GRPO2026.013344.7
WebAnchor-3BRL Algo=First-step GRPO2026.0130.144.6
WebAnchor-3BRL Algo=Anchor-GRPO2026.0128.343.5
WebAnchor-3BRL Algo=GRPO2026.0127.243.6
R1-Searcher-7BRL Algo=-2026.0120.4—