Arcee AI

Trinity Large Preview

Type
language model
Context
131K tokens
Max output
131K
Released
28 January 2026
API string
trinity-large-preview-free
Weights
open

The provider has deprecated this model. It still responds, but it is not a good choice for new projects.

Prices by provider

Providers do have offers, but we have not collected a current price yet. That is a gap on our side, not a model missing from the market.

Measurement results

Task set Result Run conditions Measured by
Arena Score, programming 1,380.41 1,373–1,388
Arena Score in French 1,379.89 1,358–1,402
Arena Score in English 1,357.6 1,352–1,363
Arena Score, long queries 1,357.15 1,350–1,364
Arena Score, expert questions 1,351.15 1,339–1,364
Arena Score, hard prompts 1,350.52 1,345–1,356
Arena Score in Spanish 1,349.24 1,328–1,370
Arena Score, multi-turn dialogue 1,342.67 1,334–1,352
Arena Score, overall 1,338.9 1,335–1,343
Arena Score, mathematics 1,335.54 1,321–1,350
Arena Score, instruction following 1,335.54 1,329–1,342
Arena Score in Russian 1,326.66 1,315–1,338
Arena Score, creative writing 1,319.39 1,310–1,329