GenomeQA (Multiple Choice): leaderboard

GenomeQA (2026) tests general-purpose LLMs on raw DNA sequences with questions from biological databases in six task families: enhancer and promoter identification, splice sites, taxonomy, histone marks, transcription factor binding sites and motifs. This board is accuracy in percent on the four-option questions, the unweighted mean of the six task families, with thinking enabled where supported; higher is better, and chance is 25.

Metric: Accuracy (%) on the four-option (chance 25) GenomeQA questions over raw DNA sequences, unweighted mean of the six task families, one fixed system prompt, thinking enabled where supported; higher is better. Source: arxiv.org. Saturation forecast: Around February 2027. 6 models tracked. Updated 7 Oct 2026.

Top models

#ModelSettingScoreOverall rank
1Gemini 3 Pro (Preview)60.87#123
2Claude Sonnet 4.5Thinking52.8#267
3GPT-5.1Thinking52.3#248
4Grok 4.1 FastReasoning50.67#431
5Qwen 3 Max (Preview)Thinking45.33#453
6Llama 4 Maverick Instruct41.53#1126

No result here: #2 Claude Fable 5.2, #3 Gemini 4 Neon, #4 GPT-6.2 Astra.

Interactive version: theaggregate.ai/benchmark?slug=genomeqa-multiple-choice · How It Works · Data refreshed daily, snapshot 2026-11-22.