# Trinity-Large-Preview Benchmark Scores & Performance

> Trinity-Large-Preview by Arcee AI scores 55.39/100 overall, ranking #100 out of 411 AI models.

Canonical page: https://benchlm.ai/models/trinity-large-preview

Last updated: September 4, 2026

## Model Details

| Property | Value |
|----------|-------|
| Creator | Arcee AI |
| Source Type | Open Weight |
| Reasoning Type | Non-Reasoning |
| Context Window | 512K |
| Overall Score | 55.39/100 |
| Overall Rank | #100 of 411 |

## Family & Coverage

- Family: Trinity Large
- Variant: preview (preview)
- Benchmarks covered: 20 of 422
- Sibling models: [Trinity-Large-Thinking](/models/trinity-large-thinking)
- Coverage note: BenchLM currently has partial benchmark coverage for this model, so the overall score is conservative.

## Agentic Benchmarks

| Benchmark | Score |
|-----------|-------|
| [τ²-bench results](/benchmarks/tau2-bench) | 90.1% |
| [GDPval-AA](/benchmarks/gdpvalaanormalized) | 3.3% |
| [GDPval-AA](/benchmarks/gdpvalaa) | 565 |
| [AA Agentic Index](/benchmarks/aaagenticindex) | 3.7% |

## Coding Benchmarks

| Benchmark | Score |
|-----------|-------|
| [AA-SciCode](/benchmarks/aascicode) | 36.1% |
| [AA Coding Index](/benchmarks/aacodingindex) | 25.8% |

## Multimodal & Grounded Benchmarks

| Benchmark | Score |
|-----------|-------|
| [Design Arena Website](/benchmarks/designarenawebsite) | 1152 |

## Reasoning Benchmarks

| Benchmark | Score |
|-----------|-------|
| [AA-LCR](/benchmarks/lcr) | 38.3% |
| [CritPt](/benchmarks/critpt) | 0.9% |

## Knowledge Benchmarks

| Benchmark | Score |
|-----------|-------|
| [MMLU](/benchmarks/mmlu) | 87.2% |
| [MMLU-Pro (Arcee)](/benchmarks/mmluproarcee) | 75.2% |
| [GPQA-D](/benchmarks/gpqa-diamond) | 63.3% |
| [Artificial Analysis Intelligence Index](/benchmarks/artificialanalysis) | 18.7% |
| [AA-GPQA Diamond](/benchmarks/aagpqadiamond) | 75.2% |
| [AA-HLE](/benchmarks/aahle) | 15.8% |
| [AA-Omniscience Index](/benchmarks/aaomniscienceindex) | -44.1% |
| [AA-Omniscience Accuracy](/benchmarks/omniscienceaccuracy) | 22.5% |
| [AA-Omniscience Hallucination Rate](/benchmarks/omnisciencehallucinationrate) | 85.9% |

## Instruction Following Benchmarks

| Benchmark | Score |
|-----------|-------|
| [AA-IFBench](/benchmarks/aaifbench) | 56.3% |

## Mathematics Benchmarks

| Benchmark | Score |
|-----------|-------|
| [AIME25 (Arcee)](/benchmarks/aime2025arcee) | 24.0% |

## Other Arcee AI Models

- [Trinity-Large-Thinking](/models/trinity-large-thinking) - Score: 47.12