List your company

General Deep Research Tool Use on Xbench DeepSearch

76Success RateGemini-3-Pro
21.9235.965064.04Mar 10, 2026

Evaluation results

MethodLinks
Gemini-3-ProCategory=Frontier (≫8B), Temperature=1, Top-p=1, Number of runs=42026.0376
GPT-OSS-120BCategory=Frontier (≫8B), Temperature=1, Top-p=1, Number of runs=42026.0369.5
DeepSeek-V3.2-ExpCategory=Frontier (≫8B), Temperature=1, Top-p=1, Number of runs=42026.0367.2
Claude-4-SonnetCategory=Frontier (≫8B), Temperature=1, Top-p=1, Number of runs=42026.0362.2
Kimi-K2-0905Category=Frontier (≫8B), Temperature=1, Top-p=1, Number of runs=42026.0361
DIVE-8B (RL)Category=Ours, Temperature=1, Top-p=1, Number of runs=42026.0358.1
Gemini-2.5-ProCategory=Frontier (≫8B), Temperature=1, Top-p=1, Number of runs=42026.0356
WebExplorer-8BCategory=8B Baselines, Temperature=1, Top-p=1, Number of runs=42026.0353.7
DIVE-8B (SFT)Category=Ours, Temperature=1, Top-p=1, Number of runs=42026.0350.2
EnvScaler-8BCategory=8B Baselines, Temperature=1, Top-p=1, Number of runs=42026.0345.7
SWE-Dev-8BCategory=8B Baselines, Temperature=1, Top-p=1, Number of runs=42026.0331.6
Qwen3-8B (base)Category=Ours, Temperature=1, Top-p=1, Number of runs=42026.0324