Benchmark Almanac

Trinity Large Preview

Trinity Large Preview.

Arcee AI Released January 27th, 2026 Family Trinity The lab's model card Compare with others
10
Benchmarks with a score
0
Ranked first, out of five or more models
0
In the top three
Arcee AI Jan 2026

Trinity Large Preview

Open weights Flagship

123.8 #323 of 404
Self-Reported Index, from the lab's own numbers. No Epoch Capabilities Index score yet.
  • MMLU-Pro 75.3% #109 /167
  • AIME 2025 24.4% #139 /143
  • MMLU 87.2% #38 /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
MMLU 87.2% #38 of 131 Lab model card
Vectara Hallucination Leaderboard (2025 dataset) 6.90% #31 of 104 Official leaderboard
LMArena (Elo) 1,379 #155 of 335 Official leaderboard
SimpleQA 23.9% #33 of 60 Lab model card
MMLU-Pro 75.3% #109 of 167 Lab model card
Longform Creative Writing (EQ-Bench) 35.0 #107 of 119 Official leaderboard

Math and science

BenchmarkScoreRankSource
GPQA Diamond 63.3% #213 of 302 Lab model card
AIME 2025 24.4% #139 of 143 Lab model card

Agents and tools

BenchmarkScoreRankSource
PinchBench 2.0 53% #47 of 57 Official leaderboard
PinchBench 40.7% #45 of 48 Official leaderboard

Its lab's numbers put it at 123.8, number 323 of 404

The Self-Reported Index fits one score to the 5 capability results Arcee AI 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 119.5 to 128.2. How we compute it.

Its launch post

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

Its system card

Every result in the longer documents the lab published with it. 1 of them isn't in the launch post.