readable AI benchmarks (simplified)

readable AI benchmarks

Trillion Labs

Tri-21B-think Preview

reasoning · open weights · Feb 10, 2026

Tri-21B-think Preview has a Quality Score of 16.8, ranking 447th among 551 scored models. It has a 32k token context window. Its Reliability Score is 18.9, ranking 472nd among 553 scored models, below average (total average: 35.6).

Quality Score16.8coverage 100.0%: AA Intelligence and AA-Omniscience outcomes available
Value Scorenot available
Factual reliability18.9
Cache Discountnot available

Model specification

Reasoning
reasoning
Input modalities
none
Output modalities
none
Context window
32k tokens
Weights
open weights

Token prices USD per 1M tokens

Input
not available
Output
not available
Cached Input
not available
Cache write
not available
Cost per task
not available

Capability

Intelligence
9.6
Coding
not available
Agentic
not available
Omniscience
-62.3
Correct
10.1%
Blended price
not available

Answer outcomes

Fully graded outcomes (Correct + Incorrect): 82.5%

Correct 10.1%Incorrect 72.4%Partial / not attempted 17.5%

Artificial Analysis benchmarks

Humanity’s Last Exam
6%
GPQA Diamond
54%
CritPt
0%
AA-Omniscience
19%
AA-LCR
22%

Similar models

Support me! Patreon