LLM Reference

Trinity-Large-Thinking

trinity-large-thinking

Researched 32d ago

Last refreshed 2026-05-11. Next refresh: weekly.

Open SourceRAGAgentsLong contextClassificationJSON / Tool use

Trinity-Large-Thinking is worth evaluating for rag, agents, and long context when its provider route and context window match the workload.

Decision context: RAG task fit, 2 tracked provider routes, and research from 2026-04-19.

Use it for

  • Teams evaluating rag, agents, and long context
  • Workloads that can use a 256K context window
  • Buyers comparing 2 tracked provider routes

Do not use it for

  • Vision or document-understanding workloads

Cheapest output

$0.850

OpenRouter per 1M tokens

Provider routes

2

Tracked API hosts

Quality / dollar

Unknown

No task benchmark coverage yet

Freshness

2026-04-19

Researched 32d ago

aging

Top use-case fit

RAG

Included by capability and metadata signals in the decision map.

Agents

Included by capability and metadata signals in the decision map.

Long context

Included by capability and metadata signals in the decision map.

Provider price ladder

Compare all 2
ProviderInput / 1MOutput / 1MRoute
OpenRouter$0.220$0.850
Serverless
Arcee AI--
Partial

Benchmark peer barsfor RAG

No task-mapped benchmark peers are available for this model yet.

Migration checks

No linked migration route is available for this model yet.

About

Arcee AI's flagship 400B sparse MoE reasoning model with 13B active parameters per token. Trained on 20T tokens with a STEM-focused curriculum. Designed for agentic workflows, chain-of-thought reasoning, and long-context tasks up to 256K tokens (BF16 API). Open-source under Apache 2.0. Available via Arcee AI API.

Trinity-Large-Thinking has a 256K-token context window.

Trinity-Large-Thinking input tokens at $0.22/1M, output at $0.85/1M.

Capabilities

ReasoningFunction CallingTool UseStructured Outputs

Benchmark Scores(1)

Scores are benchmark-specific and are direction-aware: the same numeric gap can mean very different outcomes across suites. Use the leaderboard context and this model's provider route to decide whether the winning margin is meaningful for your workload.
BenchmarkScoreVersionSource
Google-Proof Q&A89.2diamondhttps://docs.arcee.ai/language-models/trinity-large-thinking

Rankings

Show all 28 popular comparisonssorted by 7-day search impressions

Specifications

FamilyTrinity
Released2026-04-01
Parameters400B
Context256K
ArchitectureSparse Mixture of Experts (MoE)
Specializationreasoning
LicenseApache 2.0
Trainingfinetuned

Created by

Agentic AI workflows, efficient and secure

San Francisco, California, United States
Founded 2023
Website

Providers(2)