Trinity Large Thinking: benchmark scores, price and ranking
Trinity Large Thinking from Arcee AI ranks #684 of 743 LLMs on the LLMs Tiger aggregate benchmark score (43.2), based on 4 benchmark results captured up to 2026-10-13.
- Organisation
- Arcee AI
- Type
- Open-weight / local model
- Open weights
- Yes
- Cheapest API price
- $0.250 in / $0.800 out per million tokens (via openrouter, observed 2026-10-13T20:47Z)
- Aggregate rank
- #684 of 743 (score 43.2)
Relative to other models, Trinity Large Thinking is strongest on CritPt (#71 of 115) and weakest on Surface Evolver (#13 of 13).
Across the 4 benchmarks where at least ten models are scored, Trinity Large Thinking is in the top 10% on 0 and in the top half on 0.
Its cheapest listed input price, $0.250 per million tokens, is lower than that of 64% of the 427 models with a listed price.
Trinity Large Thinking benchmark results
| Benchmark | Score | Rank | Source | Captured |
|---|---|---|---|---|
| CritPt | 0.9% | #71 of 115 | Epoch AI Benchmarking Hub | 2026-10-13 |
| SciCode | 36.1% | #76 of 106 | Epoch AI Benchmarking Hub | 2026-10-13 |
| WebDev Arena | 1,237 | #61 of 67 | Epoch AI Benchmarking Hub | 2026-10-13 |
| Surface Evolver | 15.6% | #13 of 13 | Epoch AI Benchmarking Hub | 2026-10-13 |
About Trinity Large Thinking
Trinity Large Thinking is an open-weight model that can run on your own hardware from Arcee AI. The cheapest API listing we track charges $0.250 per million input tokens and $0.800 per million output tokens, through openrouter. We hold 4 benchmark results for it across 3 areas, and none is in the top tenth of its board.
Where Trinity Large Thinking stands
On our aggregate score, Trinity Large Thinking is #684 of 743 ranked models, in the lower half. Its score is 43.2; the scale is centred on 50, so it sits 6.8 points below the typical model. That beats 8% of the ranked field.
Trinity Large Thinking, benchmark by benchmark
Trinity Large Thinking lands below the middle on CritPt: 0.9%, #71 of 115, with 61% of the field scoring higher. GPT-5.6 Sol (max) leads at 32.3%; the closest result above is Qwen3.6 27B at 0.9%. It covers unpublished research-level physics problems. The board median is 2.9%, so Trinity Large Thinking is 2.0 points below it. Source: Epoch AI Benchmarking Hub, captured 2026-10-13.
Trinity Large Thinking lands below the middle on SciCode: 36.1%, #76 of 106, with 71% of the field scoring higher. Claude Opus 5.5 (max) leads at 66.9%; the closest result above is Mistral Large 3 at 36.2%. It covers scientific research code. The board median is 44.3%, so Trinity Large Thinking is 8.2 points below it. Source: Epoch AI Benchmarking Hub, captured 2026-10-13.
WebDev Arena is a weak spot. Trinity Large Thinking scores 1,237, #61 of 67, behind 90% of the models measured. The leader, Claude Opus 5.5 (max), reaches 1,820. It covers building web apps, crowd-voted (Elo). The board median is 1,433, so Trinity Large Thinking is 197 points below it. Source: Epoch AI Benchmarking Hub, captured 2026-10-13.
Surface Evolver is a weak spot. Trinity Large Thinking scores 15.6%, #13 of 13, behind 92% of the models measured. The leader, Kimi K3 (max), reaches 93.0%. It covers geometry optimisation with a solver. The board median is 46.2%, so Trinity Large Thinking is 30.6 points below it. Source: Epoch AI Benchmarking Hub, captured 2026-10-13.
Strengths and weaknesses
Software engineering: weak, ahead of 19% of the field on average across 2 benchmarks. Its best board here is SciCode (#76 of 106) and its weakest is WebDev Arena (#61 of 67).
Price and value
The median input price across the 427 priced models is $0.500 per million tokens, so Trinity Large Thinking costs less than the typical listing. Models that score within 1.5 points or better and cost less: Qwen2.5-Coder-7B-Instruct ($0.010, 49.5), DeepSeek V4.1 Flash (high) ($0.010, 48.9), Qwen2.5-Coder-3B-Instruct ($0.010, 46.5). Similar scores cost more on DeepSeek-V3.2 (Thinking; Novita) ($0.259) and GPT-4.1 mini ($0.400).
Reading these numbers
The 4 scores on this page come from 1 source: Epoch AI Benchmarking Hub (4). No official leaderboard supplied any of them, so treat them as third-party measurements. With 4 results the aggregate rank can move a lot when one new score arrives.
Models ranked nearby
- #681 granite-3.0-2b-base (43.3)
- #682 open-mixtral-8x7b (43.2)
- #683 SWE-Llama 7B (43.2)
- #685 granite-4.0-micro (43.2)
- #686 SWE-Llama 13B (43.2)
- #687 claude-3-opus-20240229 (43.2)
Compare it on the live leaderboard → · How the aggregate score works