Share your thoughts, 1 month free Claude Pro on usSee more

Professional deep-research writing on Deepresearch-Gym

77.1KPR

Claude-4-sonnet

Updated 13d ago

Evaluation Results

Method	Links
Claude-4-sonnet 2026.04		77.1	1.87	7.86
GPT-5 2026.04		76.26	2.2	7.98
Gemini2.5-Pro 2026.04		73.2	2.68	7.4
Qwen3-8B + R2-Write-SFT + RLp 2026.04		72.5	2.1	7.8
Qwen3-8B + R2-Write-SFT + RL 2026.04		71.63	2.24	7.46
Qwen3-8B + R2-Write-SFT 2026.04		70.05	2.33	7.19
Qwen3-4B + R2-Write-SFT + RLp 2026.04		69.72	2.42	7.1
Qwen3-8B + Distill + RL 2026.04		68.92	2.7	7.25
Qwen3-8B + RL 2026.04		68.84	2.73	6.86
LongWriter-Zero-32B 2026.04		68.2	3.24	7.05
Writing-RL-7B 2026.04		68.1	3.62	7.05
Qwen3-4B + R2-Write-SFT 2026.04		67.12	2.68	6.44
Qwen3-8B + Distill 2026.04		66.37	2.48	6.6
Qwen3-8B + R2-Write-Last 2026.04		65.72	2.62	6.6
Qwen3-4B + RL 2026.04		64.2	2.84	6.23
Qwen3-8B 2026.04		63.78	2.62	5.98
Qwen3-4B 2026.04		60.04	2.88	5.76
Reverse-Engineering-8B 2026.04		60.01	4.89	6.02
Longwriter-9B 2026.04		57.22	5.98	5.66