PhageBench - Completeness Estimation: leaderboard

Grading a phage genome as complete or of high, medium or low quality from raw bacteriophage genome sequences on PhageBench (2026), multiple-choice questions built from curated phage databases and answered zero-shot with chain-of-thought prompting. Accuracy in percent; higher is better.

Metric: Accuracy (%) on 1,000 four-option items grading a phage genome as complete, high, medium or low quality, each a raw-nucleotide multiple-choice question with shuffled options, zero-shot chain-of-thought prompting (reasoning models at reasoning effort medium with at most 2,048 reasoning tokens); higher is better. Source: arxiv.org. Saturation forecast: Around June 2028. 8 models tracked. Updated 7 Oct 2026.

Top models

#ModelSettingScoreOverall rank
1GPT-5.2Medium43.1#185
2Claude Sonnet 4.5Thinking35.5#261
3Qwen 3 MaxThinking33.8#415
4Gemini 3 FlashMedium32.1#175
5LLaMA 4 (PhageBench checkpoint unspecified)27.1
6Qwen3-235B (PhageBench checkpoint unspecified)25.4
7GPT-4o Mini24.5#1667
8GPT-OSS-120B24.2#704

No result here: #1 Claude Fable 5.2, #3 Gemini 4 Neon, #4 GPT-6.2 Astra.

Interactive version: theaggregate.ai/benchmark?slug=phagebench-completeness-estimation · How It Works · Data refreshed daily, snapshot 2026-11-21.