List your company

Automated Peer Review on DeepReview-13K 2025 (test)

98.7Technical Accuracy WinCycleReviewer-8B
30.89248.49666.183.704Jan 30, 2026

Evaluation results

MethodLinks
CycleReviewer-8BCategory=Fine-tuned, Size=8B2026.0198.70.91000100098.50.399.70.2
CycleReviewer-70BCategory=Fine-tuned, Size=70B2026.0198.7199.80.299.8098.60.499.50.4
DeepReviewer-7BCategory=Fine-tuned, Size=7B2026.0193.24.298.11.797.61.3961.797.62.2
DeepReviewer-14BCategory=Fine-tuned, Size=14B2026.0177.911.789.29.385.910.990.34.891.58.3
AI Scientist v2 (Gemini 3 Pro)Category=Multi Agent, Backbone=Gemini 3 Pro2026.0149.225.883.813.282.311.47810.781.618
Agent Review (Gemini 3 Pro)Category=Multi Agent, Backbone=Gemini 3 Pro2026.0145.328.283.61380.113.47910.18018.9
Claude Sonnet 4.5Category=Single Agent2026.0144.724.474.120.353.836.578.212.775.123.4
Gemini 3 ProCategory=Single Agent2026.0140.23376.918.169.122.574.911.87326.1
Gemini 3 FlashCategory=Single Agent2026.0139.532.481.616.774.918.777.91378.321.7
AI Scientist v2 (Gemini 3 Flash)Category=Multi Agent, Backbone=Gemini 3 Flash2026.0138.134.27620.463.725.871.215.57127
AI Scientist v2 (Claude Sonnet 4.5)Category=Multi Agent, Backbone=Claude Sonnet 4.52026.0137.226.169.826.644.746.772.917.165.834.2
Stanford Agent ReviewerCategory=Multi Agent, Access=Browser based, Sample Size=50 papers sampled from DeepReview-13K2026.0136325246384660345446
Agent Review (Gemini 3 Flash)Category=Multi Agent, Backbone=Gemini 3 Flash2026.0134.727.877.120.367.626.975.413.476.523
Agent Review (Claude Sonnet 4.5)Category=Multi Agent, Backbone=Claude Sonnet 4.52026.0133.53359.438.632.555.869.116.257.941.6