BenchLeader
Arcee AIAuto-detected

Trinity Large

Best configuration ranks #318 of 610 on the BenchLeader Index at 48.7 ±5.0 (thinking reasoning effort). Last measured 8 Sept 2026. Released 1 Apr 2026.

Blended price
$0.413/M
$0.250 in · $0.900 out
Output speed
321 tok/s
First answer
7.20 s
first token 0.21 s
Context
512k
How it scores by categoryDashed line = average model (50). One step of 15 = one standard deviation.
  1. Overall index49
  2. Reasoning55
  3. Coding37
  4. Agents & tools53
  5. Knowledge44
  6. Instruction following56
  7. Human preference55
  8. Long context45
  9. Composite43

Reasoning-effort configurations

The same model behaves differently depending on how much it is allowed to think. Each row is one setting, scored only on the benchmarks that were run at that setting. “Default” means the publisher did not say which setting was used.

EffortIndexRankSpeedFirst answerChat reply costCategories
thinkingbest48.7#318321 tok/s7.20 s$0.0004Agents & tools 53 · Coding 37 · Composite 43 · Human preference 55 · Instruction following 56 · Knowledge 44 · Long context 45 · Reasoning 55
default105 tok/s0.21 sCoding 58 · Human preference 56 · Reasoning 57

Benchmark results

One column per reasoning effort. Rank is among every configuration of every model on that benchmark. Hover a score for the run it came from.

Reasoning

BenchmarkthinkingdefaultSourceTrend
LMArena Hard Prompts1386#1591404#149LMArena
GPQA Diamond (AA)not in index75.2%#242Artificial Analysis
Humanity's Last Exam (AA)not in index15.8%#190Artificial Analysis

Coding

BenchmarkthinkingdefaultSourceTrend
SciCode36.1%#125SciCode
LMArena Coding1414#1621442#136LMArena
LMArena WebDev1237#110LMArena
SciCode (AA)not in index40.6%#112Artificial Analysis

Agents & tools

BenchmarkthinkingdefaultSourceTrend
Terminal-Bench Hard22.7%#146Artificial Analysis
τ²-Bench Telecom (AA)not in index90.1%#66Artificial Analysis

Knowledge

BenchmarkthinkingdefaultSourceTrend
AA-Omniscience-44.1#281Artificial Analysis
CorpFin54.7%#90Vals AI

Instruction following

BenchmarkthinkingdefaultSourceTrend
IFBench56.3%#141Artificial Analysis

Human preference

BenchmarkthinkingdefaultSourceTrend
LMArena Text1369#1611379#153LMArena

Long context

BenchmarkthinkingdefaultSourceTrend
AA-LCR38.0%#294Artificial Analysis

Composite

BenchmarkthinkingdefaultSourceTrend
AA Intelligence Index10.9#300Artificial Analysis

Where to run it

Every provider serving this model through OpenRouter, with throughput and first-token latency measured on live traffic over the last 30 minutes and each provider’s own price. Purple marks the best in each column.

ProviderSpeedFirst tokenInput $/MOutput $/MContextQuantisation
Arcee AI105 tok/s0.21 s$0.250$0.800262k

What a task costs

Estimates from list price, output speed and time to first answer for the best configuration. “With caching” assumes three-quarters of the input is served from the prompt cache. Reasoning tokens are not modelled.

WorkloadTokens in / outCostWith cachingTime
Chat reply400 / 300$0.00048.1 s
Summarise a 30-page report12,000 / 600$0.00359.1 s
Code edit6,000 / 1,500$0.002911.9 s
Agentic coding session60,000 / 4,000$0.01919.7 s
Structured extraction2,000 / 200$0.00077.8 s

See also

Data as of 9 Sept 2026. Compare these configurations.