Noometry

Arcee AI, open weights

Trinity Large Thinking

Trinity Large Thinking by Arcee AI ranks 185th of 354 ranked models on the Noometry Index as of October 2026, with a score of 38.6. Its strongest category is knowledge, where it ranks 113th. API pricing starts at $0.25 per million input tokens and $0.80 per million output tokens, with a 262K-token context window.

Last verified

Noometry rank
#185 of 354
Index score
38.6
Evidence
Confirmed 24 results
Provider
Arcee AI
Released
April 1, 2026
Weights
Open weights
Reasoning
Yes
Context window
262K
Max output
80K
Input price
$0.25 / M
Output price
$0.80 / M
Blended price
$0.39 / M
Output speed
Not measured
Value
#61 of 219
Knowledge cutoff
Unknown
Input
text
Hugging Face
arcee-ai/Trinity-Large-Thinking

Category scores

Each category score combines every public result we have in that category.

  1. Coding 34.1
  2. Reasoning 16.9
  3. Math 37.6
  4. Knowledge 40.9
  5. Multilingual 46.2
  6. Instruction Following 70.5
  7. Long Context 41.3
  8. Writing & Preference 53.8
CategoryScoreRankResults
Coding34.1#2443
Reasoning16.9#2985
Math37.6#1491
Knowledge40.9#1132
Multilingual46.2#1601
Instruction Following70.5#1621
Long Context41.3#1441
Writing & Preference53.8#1583

Strengths and weaknesses

Categories where Trinity Large Thinking places highest and lowest among the models ranked in each, with its score against that category's median.

Strongest categories

CategoryScorevs medianRank
Knowledge40.9+3.6#113 of 314, top 36%
Math37.6+1.1#149 of 327, top 46%
Long Context41.3+0.3#144 of 296, top 49%

Weakest categories

CategoryScorevs medianRank
Reasoning16.9−6.7#298 of 350, top 86%
Coding34.1−4.6#244 of 340, top 72%
Multilingual46.2−1.2#160 of 297, top 54%

Closest competitors

The models ranked just above and below Trinity Large Thinking. When scores are this close, price and speed are often the better way to choose.

ModelRankScoreBlended $/MSpeed
Qwen2.5 Plus 1127#18138.8——Compare
Qwen3.6 Flash#18238.8$0.42—Compare
Olmo 3 32b Think#18338.7——Compare
Hunyuan Large 2025 02 10#18438.6——Compare
GPT-5.1-Codex#18638.6$3.44—Compare
Sonar#18738.5$1—Compare
MiniMax-M2.5#18838.3$0.5249Compare
Nova Premier 1.0#18938.3$59Compare

Sponsored placements are available on pages like this one. Advertise on Noometry

Benchmark results

Every published result we track, with its source. Bold rows are the ones used for ranking; where several exist we prefer independent runs over self-reported numbers.

Coding

BenchmarkScorePositionSettingSourceDate
LMArena WebDev1238#105 of 113, top 93%thinkingLMArena2026-10-08
SciCode36.1%#92 of 121, top 77%Epoch AI
LMArena Coding1381#147 of 294, top 50%LMArena2026-10-08
LMArena Coding1357thinkingLMArena2026-10-08

Reasoning

BenchmarkScorePositionSettingSourceDate
NYT Connections (extended)16.5%#81 of 91, top 90%Lech Mazur benchmarks
CritPt0.9%#86 of 134, top 65%Epoch AI
Thematic Generalization41.6%#20 of 23, top 87%Lech Mazur benchmarks
LMArena Hard Prompts1350#160 of 297, top 54%LMArena2026-10-08
LMArena Hard Prompts1342thinkingLMArena2026-10-08
Surface Evolver Bench15.6%#25 of 25, top 100%Epoch AI

Math

BenchmarkScorePositionSettingSourceDate
LMArena Math1336LMArena2026-10-08
LMArena Math1366#152 of 285, top 54%thinkingLMArena2026-10-08

Knowledge

BenchmarkScorePositionSettingSourceDate
Vectara Hallucination Rate (lower is better)6.9%#27 of 96, top 29%Vectara Hallucination Leaderboard
LMArena Expert1352LMArena2026-10-08
LMArena Expert1360#148 of 273, top 55%thinkingLMArena2026-10-08

Multilingual

BenchmarkScorePositionSettingSourceDate
LMArena Non-English1318LMArena2026-10-08
LMArena Non-English1325#160 of 297, top 54%thinkingLMArena2026-10-08
LMArena Chinese1352LMArena2026-10-08
LMArena Chinese1373#155 of 285, top 55%thinkingLMArena2026-10-08
LMArena French1374#133 of 223, top 60%LMArena2026-10-08
LMArena French1365thinkingLMArena2026-10-08
LMArena German1313LMArena2026-10-08
LMArena German1356#124 of 231, top 54%thinkingLMArena2026-10-08
LMArena Japanese1290LMArena2026-10-08
LMArena Japanese1311#115 of 211, top 55%thinkingLMArena2026-10-08
LMArena Korean1258LMArena2026-10-08
LMArena Korean1306#128 of 213, top 61%thinkingLMArena2026-10-08
LMArena Russian1325LMArena2026-10-08
LMArena Russian1337#153 of 283, top 55%thinkingLMArena2026-10-08
LMArena Spanish1357#139 of 226, top 62%LMArena2026-10-08
LMArena Spanish1328thinkingLMArena2026-10-08

Instruction Following

BenchmarkScorePositionSettingSourceDate
LMArena Instruction Following1334#155 of 298, top 53%LMArena2026-10-08
LMArena Instruction Following1325thinkingLMArena2026-10-08

Long Context

BenchmarkScorePositionSettingSourceDate
LMArena Longer Query1355#150 of 291, top 52%LMArena2026-10-08
LMArena Longer Query1318thinkingLMArena2026-10-08

Writing & Preference

BenchmarkScorePositionSettingSourceDate
LMArena Text1339LMArena2026-10-08
LMArena Text1340#164 of 297, top 56%thinkingLMArena2026-10-08
LMArena Creative Writing1320#152 of 295, top 52%LMArena2026-10-08
LMArena Creative Writing1307thinkingLMArena2026-10-08
LMArena Multi-Turn1342#158 of 295, top 54%LMArena2026-10-08
LMArena Multi-Turn1320thinkingLMArena2026-10-08

API pricing by provider

RouteInput $/MOutput $/MCached input $/MChecked
openrouter$0.25$0.80$0.062026-10-10

All Arcee AI API prices → Estimate your cost →

Compare Trinity Large Thinking

Frequently asked questions

How good is Trinity Large Thinking?

Trinity Large Thinking by Arcee AI ranks 185th of 354 ranked models on the Noometry Index as of October 2026, with a score of 38.6. Its strongest category is knowledge, where it ranks 113th. API pricing starts at $0.25 per million input tokens and $0.80 per million output tokens, with a 262K-token context window.

How much does Trinity Large Thinking cost?

Trinity Large Thinking costs $0.25 per million input tokens and $0.80 per million output tokens on openrouter, with cached input at $0.06.

What is Trinity Large Thinking's context window?

Trinity Large Thinking accepts up to 262K tokens of input and can write up to 80K tokens in one response.

Is Trinity Large Thinking open source?

Yes. Trinity Large Thinking's weights are downloadable from Hugging Face (arcee-ai/Trinity-Large-Thinking); check the license for commercial terms.

What are Trinity Large Thinking's strengths and weaknesses?

Relative to other ranked models, Trinity Large Thinking places best in knowledge, math, long context and lowest in reasoning, coding, multilingual.

What is Trinity Large Thinking best at?

Its best category is knowledge, where it ranks 113th on Noometry.