Benchmark Almanac

Cogito v2 Preview 671B MoE

Cogito v2 Preview 671B MoE ranks first on 2 of the 8 benchmarks with five or more models scored.

Deep Cogito Released July 31st, 2025 Family Cogito v2 The lab's model card Compare with others
8
Benchmarks with a score
2
Ranked first, out of five or more models
2
In the top three
Deep Cogito Jul 2025

Cogito v2 Preview 671B MoE

Open weights Flagship

137.8 #197 of 404
Self-Reported Index, from the lab's own numbers. No Epoch Capabilities Index score yet.
  • MMLU-Pro 85.4% #22 /167
  • MMMLU 82.4% #39 /71
  • MMLU 91.3% #4 /131

Where it ranks

Its best score on each benchmark, and where that ranks among every model with a score there. We rank each model once, at its best setting.

Knowledge and language

BenchmarkScoreRankSource
MGSM 93% #1 of 57 Lab model card
MMLU 91.3% #4 of 131 Lab model card
MMLU-Pro 85.4% #22 of 167 Lab model card
SimpleQA 31.4% #21 of 60 Lab model card
MMMLU 82.4% #39 of 71 Lab model card

Math and science

BenchmarkScoreRankSource
MATH 98.2% #1 of 99 Lab model card
GSM8K 96.2% #6 of 92 Lab model card
GPQA Diamond 76% #144 of 302 Lab model card

Its lab's numbers put it at 137.8, number 197 of 404

The Self-Reported Index fits one score to the 8 capability results Deep Cogito reported for this model, on the scale of the Epoch Capabilities Index. Labs choose which benchmarks to show and how to run them, so read it as the lab's claim, not an independent measurement. The 90% interval runs from 134.8 to 140.8. How we compute it.

Its launch post

The benchmarks the lab chose to headline, with the scores it gave.