Model profile
Trinity Large
Evidence summary
Safety. Trinity Large has an estimated Safety rank of #263; its 90% source-sensitivity interval is #111–#290. Its behavior-only rank is #266; company governance moves the combined estimate to #263. Published Safety evidence spans 7 eval lineages and 4 of 7 components. Its strongest relative result is Arena Factuality — Text Arena (factuality-only weighting) (factuality_bt_rating, #27 of 111); its weakest is SM-Bench (ambiguous_interpretation, #82 of 84).
Freedom. Trinity Large has an estimated Freedom rank of #84; its 90% source-sensitivity interval is #34–#348. Published Freedom evidence spans 3 eval lineages and 1 of 1 components. Its strongest relative result is SM-Bench (adversarial, #8 of 84); its weakest is SM-Bench (overfit, #54 of 84).
Compare this model
Only models sharing at least one published sub-eval are listed.
Official and reference links
- Artificial Analysis ↗trinity-large-thinking
- Hugging Face ↗arcee-ai/Trinity-Large-Preview
- OpenRouter ↗arcee-ai/trinity-large-preview
- Official model page ↗Exact model document · Reviewed official Hugging Face owner · official repository
- Release source ↗repository created at proxy
Safety evals
Rank and direction are specific to the Safety portfolio. Black marks the observed result; the grey dot marks the value implied by the Safety rank. Values stay on each source’s native scale.
| Eval / sub-eval | Rank | Value | Distribution | Source |
|---|---|---|---|---|
| AA-Omnisciencehallucination_rate | #225 / 345 | ↓0.8589 | Source ↗official | |
| Arena Factuality — Text Arena (factuality-only weighting)factuality_bt_rating | #27 / 111 | ↑1451.0 | Source ↗official | |
| BullshitBench v2clear_pushback_rate | #85 / 117 | ↑0.15 | Source ↗official | |
| DystopiaBenchbasaglia_score | #46 / 50 | ↓71.4 | Source ↗official | |
| DystopiaBenchbaudrillard_score | #28 / 50 | ↓58.53 | Source ↗official | |
| DystopiaBenchhuxley_score | #46 / 50 | ↓80.8 | Source ↗official | |
| DystopiaBenchlaguardia_score | #47 / 50 | ↓73.17 | Source ↗official | |
| DystopiaBenchorwell_score | #44 / 50 | ↓75.1 | Source ↗official | |
| DystopiaBenchpetrov_score | #45 / 50 | ↓81.87 | Source ↗official | |
| Opposite-Narrator Sycophancysycophancy_rate_pct | #23 / 24 | ↓18.9 | Source ↗official | |
| SM-Benchadversarial | #77 / 84 | ↑73.42 | Source ↗official | |
| SM-Benchambiguous_interpretation | #82 / 84 | ↑62.05 | Source ↗official | |
| SM-Benchanti_hallucination | #71 / 84 | ↑80.89 | Source ↗official | |
| SM-Bencheq_boundaries | #36 / 84 | ↑66.86 | Source ↗official | |
| SM-Benchoverfit | #54 / 84 | ↑65.17 | Source ↗official | |
| Vectara HHEM Factual Consistencyfactual_consistency_rate | #26 / 94 | ↑93.1 | Source ↗official |
Freedom evals
Rank and direction are specific to the Freedom portfolio. Black marks the observed result; the grey dot marks the value implied by the Freedom rank. Values stay on each source’s native scale.
| Eval / sub-eval | Rank | Value | Distribution | Source |
|---|---|---|---|---|
| SM-Benchadversarial | #8 / 84 | ↓73.42 | Source ↗official | |
| SM-Bencheq_boundaries | #36 / 84 | ↑66.86 | Source ↗official | |
| SM-Benchoverfit | #54 / 84 | ↑65.17 | Source ↗official | |
| SpeechMap model completioncomplete_pct | #71 / 181 | ↑65.5 | Source ↗official | |
| UGI Leaderboard — base-model willingnesswillingness_adherence_score | #25 / 156 | ↑6 | Source ↗official | |
| UGI Leaderboard — base-model willingnesswillingness_direct_score | #21 / 156 | ↑5 | Source ↗official |
Values evaluations
Descriptive values and political-framing results are separate from safety/ethics ranks. Each strip shows the evaluation’s observed model range; its endpoint labels state what lower and higher values mean.