arcee-ai Open weights Reasoning

Trinity Large Thinking pricing & benchmarks

Released Apr 1, 2026 · served by 2 providers

Input / 1M tokens$0.220
Output / 1M tokens$0.800
Context window262K
Cached input$0.060
Intelligence index18.2
Coding index25.8
Agentic index3.7
1M in + 300K out$0.46

What Trinity Large Thinking is

Trinity Large Thinking is a powerful open source reasoning model from the team at Arcee AI. It shows strong performance in PinchBench, agentic workloads, and reasoning tasks. Launch video: https://youtu.be/Gc82AXLa0Rg?si=4RLn6WBz33qT--B7...

Providers and prices for Trinity Large Thinking

The same model costs different amounts depending on who serves it. Prices are per 1M tokens, cheapest first. Uptime and throughput are OpenRouter's rolling measurements, not vendor claims.

ProviderInputOutputContext QuantUptime 24hThroughput
Arcee AIcheapest $0.250 $0.800 262K 99.4%
Parasail $0.220 $0.850 262K fp4 100.0%

Trinity Large Thinking benchmark results

Design Arena head-to-head results, as reported through the OpenRouter model API.

CategoryArenaRankEloWin rate
asciiart models #52 1079 37.1%
svg models #62 1066 35.2%
website models #68 1162 41.3%
3d models #69 1139 41.3%
codecategories models #72 1148 40.2%
gamedev models #74 1135 38.4%
dataviz models #78 1127 39.3%
uicomponent models #82 1081 32.6%

Cheaper models in the same class

Models scoring within 4 points of Trinity Large Thinking on the Intelligence Index, but with a lower output price.

FAQ

How much does the Trinity Large Thinking API cost?

$0.220 per 1M input tokens and $0.800 per 1M output tokens. A typical workload of 1M input + 300K output tokens costs about $0.46.

Is Trinity Large Thinking free?

No. It is a paid model starting at $0.220 per 1M input tokens, though some providers offer trial credits.

Which provider is cheapest for Trinity Large Thinking?

Arcee AI at $0.250 input / $0.800 output per 1M tokens.

Can I self-host Trinity Large Thinking?

Yes — weights are published as arcee-ai/Trinity-Large-Thinking on Hugging Face, so you can run it on your own hardware.