a

Trinity Large Thinking

Apr 2026
aTrinity Large Thinking
Compare in
on
$0.12per task?
$0.25 / $0.80per 1M in / outcache $0.06
347 tok/sfirst token in 1.2 s?
262Kcontext

Rankingin thinking mode?

Full ranking →
1GLM 5.299.11JT 35B Flash99.13GLM 4.7 Flash98.84Fable 598.54Step 3.7 Flash98.5
···
40MiMo v2.590.641Trinity Large Thinking90.141Kimi K2.7 Code90.1

Strong at

A
Tool calling#41 of 242
B
Coding#65 of 134Reasoning#143 of 336Science#153 of 312Instructions#100 of 245
C
Agents & work tasks#112 of 177Long context#203 of 314Knowledge#189 of 320

Run it yourself

estimate
Fits on✗ 8 GB (phone, laptop)✗ 16 GB✗ 24 GB GPU✗ 48 GB GPU✗ 80 GB GPU
Licenceopenmdw-1.1commercial use with conditions

Benchmarks

in Agents & Tool Use
Terminal-Bench 2.120.6#112
Terminal-Bench 4.00.5#65
Terminal-Bench Hard (version unstated)22.7#84
GDPval-AA (work tasks) 2.1335.8#112
AA-LCR (long context) 1.138#205
τ²-bench · AA90.1#41
τ³-Banking · AA5.8#111
AutomationBench-AA1.3#109
AA-Briefcase 1.1 · Elo255.1#114

FAQ

How much does Trinity Large Thinking cost?›

$0.25 per 1M input tokens and $0.80 per 1M output tokens; cached input $0.06 per 1M. A typical benchmark task costs about $0.12 at default effort.

Is Trinity Large Thinking free to use?›

There is no free API; the cheapest paid price is $0.25 per 1M input tokens. The weights are open, so you can run it yourself.

How good is Trinity Large Thinking, and what is it best at?›

#143 of 336 in Chat & Knowledge, #65 of 134 in Coding, #109 of 135 in Agents & Tool Use, #122 of 197 in Writing & Language, #116 of 135 in Vision & Documents. Strongest at: Tool calling.

Is Trinity Large Thinking better than MiMo v2.5?›

On τ²-bench · AA, Trinity Large Thinking scores 90.1 vs 90.6 — about the same. It is 184% more expensive per task.

How can I use Trinity Large Thinking?›

Through the API of Arcee AI, Featherless.

What is the context window of Trinity Large Thinking?›

262K tokens, with up to 80K output tokens.

When was Trinity Large Thinking released? Is there a newer version?›

Released 2026-04-01. It is the current version in its line.

Is Trinity Large Thinking open source?›

Yes, the weights are open (openmdw-1.1 licence).

Can I run Trinity Large Thinking locally?›

The weights are open and can be self-hosted; hardware needs are not confirmed yet.

Can Trinity Large Thinking read images?›

No, it accepts text only.

What is Trinity Large Thinking?›

Trinity Large Thinking is an open-weights language model from arcee-ai, released Apr 2026.

What are the best alternatives to Trinity Large Thinking?›

GLM 4.7 Flash (cheaper), MiMo v2.5 Pro (better), GLM 5.2 (open weights).