Skip to content
Model Pareto

Trinity Large Thinking

Arcee AIOpen weightsartificialanalysis.ai ↗
Input
$0.25/M tok
Output
$0.80/M tok
Blended
$0.388/M tok · 3:1
Speed
337tok/s
Context
512Ktokens
Cost per task
~$0.34/taskest.

Long-tail import from Artificial Analysis (2026-09-24); AA variant: Trinity Large Thinking. Price = AA median host price.

Rankings

Self-hostingEstimated

How we estimate →

What it would cost to run these open weights yourself on rented GPUs, compared with the API price above.

Hardware
8× H200 141GB
FP8 weights
Throughput
~29,038 tok/s
many requests batched
Price
$0.12–0.99 /M tok
busy → light use
On your own machine
Mac Studio M5 Ultra 512GB (int4, ~130 tok/s single-stream)
single consumer GPU or Mac
Assumptions (5)
  • FP8 weights (398B params, 13B active per token) + 40% KV-cache headroom ≈ 557 GB
  • 8× H200 141GB at $3.59–$7.91/GPU-hour on-demand (2026-09-24)
  • ~29,038 output tok/s aggregate at batch 1044 (bandwidth-bound); MoE compute scales with active params
  • Blended 3:1 input:output; prefill ~287,716 tok/s
  • Low = 75% utilization at the low GPU price; high = 20% at the high price

Benchmark scores

~ italic, dashed = no published score yet, estimated from related benchmarks.

Coding

~#25 of 26 (estimated)~28.6
  1. #1Claude Opus 5.561.7
  2. #2Claude Opus 559.7
  3. #24GPT-6 Luna31.1
  4. ~#25Trinity Large ThinkingEstimated28.6
  5. #25Muse Spark Open26.3
No published score — estimated from related benchmarks

Writing & chat

Hard reasoning

Agents