Arcee AI

Small merged models

Tier

Moderate

Headquarters

San Francisco, CA, USA

Founded

2023

Flagship

Arcee

Models tracked

1

Artificial Analysis

Top intelligence

31.9

Artificial Analysis Index

Cheapest model

$0.40

blended /1M

Fastest output

221

tokens / sec

Flagship

Arcee AI’s top-scoring model on the Artificial Analysis Intelligence Index.

Most intelligent model

Trinity Large Thinking

Released 2026-04-01

Intelligence

31.9

Coding

27.2

Math

In $/1M

$0.23

Out $/1M

$0.88

Speed t/s

221

Benchmark breakdown

GPQA Diamond
75
Humanity’s Last Exam
15
SciCode
36
τ²-Bench (agentic)
90
Terminal-Bench Hard
23
IFBench
56

Models & benchmarks

Every Arcee AI model on file — intelligence, coding, price, and speed.

ModelReleasedIntelCodingMathIn $/1MOut $/1MSpeed t/sLatency s
Trinity Large Thinking2026-04-0131.927.2$0.23$0.882210.77
Via Artificial Analysis

House strengths

Where Arcee AI’s lineup over-indexes, averaged across its models.

Overview

A company specializing in small, domain-adapted language models built through model merging and distillation.

Visit Arcee AI