List your company

Deep Research Report Generation on DRB

50.7Overall ScoreGPT-5 + Search
6.08417.66729.2540.833Jun 15, 2026

Evaluation results

MethodLinks
GPT-5 + SearchBackbone=GPT-52026.0650.749.751.551.648.5
Gemini Deep ResearchBackbone=Proprietary2026.0648.848.548.549.249.4
OpenAI Deep ResearchBackbone=o3/o42026.0646.946.845.249.247.1
Gemini 3 Pro + SearchBackbone=Gemini 3 Pro2026.0646.343.444.949.849
DEEPRUBRIC (140-step, ≈750h)Backbone=Qwen3-8B, Steps=140-step, Compute Cost=≈750h2026.0643.641.741.848.442.5
DR Tulu-8B (1900-step, ≈9700h)Backbone=Qwen3-8B, Steps=1900-step, Compute Cost=≈9700h2026.0643.441.741.848.241.3
Perplexity Deep ResearchBackbone=Proprietary2026.0642.340.739.346.444.3
DEEPRUBRIC (75-step, ≈402h)Backbone=Qwen3-8B, Steps=75-step, Compute Cost=≈402h2026.0641.939.74046.941.5
DR Tulu-8B (1000-step, ≈6000h)Backbone=Qwen3-8B, Steps=1000-step, Compute Cost=≈6000h2026.0641.839.539.447.341.6
Perplexity-Sonar (High)Backbone=Proprietary2026.0640.737.436.145.744.7
Tongyi DeepResearch-30B-A3BBackbone=Qwen2.5-30B2026.0640.639.134.346.845.4
WebThinker-32B-DPO (report)Backbone=Qwen2.5-32B (report)2026.0640.639.435.44643.5
QwQ-32B + RAGBackbone=QwQ-32B2026.0640.338.134.84744.6
DR Tulu-8B (SFT)Backbone=Qwen3-8B2026.063936.335.345.539.5
DEEPRUBRIC (SFT)Backbone=Qwen3-8B, Stage=SFT2026.063833.83643.540.1
WebThinker (report)Backbone=QwQ-32B (report)2026.0637.936.232.643.242.9
WebExplorer-8BBackbone=Qwen2.5-8B2026.0636.733.728.545.742.2
Ai2 ScholarQABackbone=Claude Sonnet2026.0636.135.13240.538.9
Claude-Sonnet SearchBackbone=Claude Sonnet2026.0634.53937.745.841.5
Qwen3-8B + RAGBackbone=Qwen3-8B2026.0633.329.42740.241.1
WebThinker-32B-DPOBackbone=Qwen2.5-32B2026.0623.319.712.336.826.3
Qwen3-8B + SearchBackbone=Qwen3-8B2026.0618.214.38.729.524.4
Search-R1-7BBackbone=Qwen2.5-7B2026.069.55.22.118.616.8
ASearcher-Web-7BBackbone=Qwen2.5-7B2026.067.85.11.715.211.8