Arcee AILanguage model
Trinity Large Preview
Trinity Large Preview has no overall index value yet because there are too few independent benchmark results.
35out of 100same as last month
based on search, usage and attention
Summary
Trinity Large Preview combines 398 billion total parameters with about 13 billion active and a context extended to 512K. Arcee trained the sparse MoE on more than 17 trillion tokens, made it chat ready, and continues reinforcement learning according to the model card. OpenMDW-1.1 and the Preview label make the checkpoint open but changing.
All benchmark results
Every published result with its source, test setup and comparison models.
Show all 5 published benchmark results
Measurements without matching peer values
These measurements have no matching peer values under the same test conditions. Their original values and sources remain available here.
MMLU (Show measurement, test conditions, and source)
- Source value
- 87.2
- Score
- 87.2
- Metric
- score as published
- Unit
- No unit provided
- Benchmark version
- MMLU; source table does not state task snapshot
- Category
- general
- Direction
- Higher is better
- Source type
- vendor-reported
- Evaluator
- Arcee Trinity Large Preview model-card evaluation table
- Status
- active
- Retrieved at
- 2026-10-15
- Methodology
- Vendor-reported result. Exact Trinity Large Preview column; the model card does not state prompting, sampling settings or run dates for these rows.
- Context
- Values come from the model publisher; their methods are not fully specified in the card.
MMLU-Pro (Show measurement, test conditions, and source)
- Source value
- 75.2
- Score
- 75.2
- Metric
- score as published
- Unit
- No unit provided
- Benchmark version
- MMLU-Pro; source table does not state task snapshot
- Category
- general
- Direction
- Higher is better
- Source type
- vendor-reported
- Evaluator
- Arcee Trinity Large Preview model-card evaluation table
- Status
- active
- Retrieved at
- 2026-10-15
- Methodology
- Vendor-reported result. Exact Trinity Large Preview column; the model card does not state prompting, sampling settings or run dates for these rows.
- Context
- Values come from the model publisher; their methods are not fully specified in the card.
GPQA-Diamond (Show measurement, test conditions, and source)
- Source value
- 63.3
- Score
- 63.3
- Metric
- score as published
- Unit
- No unit provided
- Benchmark version
- GPQA-Diamond; source table does not state task snapshot
- Category
- general
- Direction
- Higher is better
- Source type
- vendor-reported
- Evaluator
- Arcee Trinity Large Preview model-card evaluation table
- Status
- active
- Retrieved at
- 2026-10-15
- Methodology
- Vendor-reported result. Exact Trinity Large Preview column; the model card does not state prompting, sampling settings or run dates for these rows.
- Context
- Values come from the model publisher; their methods are not fully specified in the card.
AIME 2025 (Show measurement, test conditions, and source)
- Source value
- 24
- Score
- 24
- Metric
- score as published
- Unit
- No unit provided
- Benchmark version
- AIME 2025; source table does not state attempt protocol
- Category
- general
- Direction
- Higher is better
- Source type
- vendor-reported
- Evaluator
- Arcee Trinity Large Preview model-card evaluation table
- Status
- active
- Retrieved at
- 2026-10-15
- Methodology
- Vendor-reported result. Exact Trinity Large Preview column; the model card does not state prompting, sampling settings or run dates for these rows.
- Context
- Values come from the model publisher; their methods are not fully specified in the card.
SimpleQA (Show measurement, test conditions, and source)
- Source value
- 23.92
- Score
- 23.92
- Metric
- Score
- Unit
- No unit provided
- Benchmark version
- arXiv:2602.17004v1, Table 4
- Direction
- No direction published
- Source type
- vendor-reported
- Evaluator
- Arcee Trinity Large Technical Report, Table 4
- Status
- active
- Retrieved at
- 2026-10-19
- Methodology
- Vendor-reported result. Trinity Large Preview instruct model evaluation suite; Table 4 does not specify row-level prompting, decoding, dataset split, or run date.
- Context
- Table 4 labels the numeric field only “Score”. The row does not state its unit or scale, score definition, direction, sample size, confidence interval, or immutable evaluated checkpoint.
Popularity compared
How often the model is searched for, used and mentioned, compared with every model of its category.
Popularity index
The 100 most popular models of this category, 0 to 100, higher is more popular
20 of 101 models
As of November 2026 · Source: gradually.ai popularity index
Over time
Weekly values from October 18, 2026 to November 7, 2026
Trinity Large PreviewMagistral Small 1.2 35Mistral Large 2407 35Jamba Reasoning 3B 35MiMo-V2-Flash 35
Method changes in the timeline, a jump at these points is no market event:
- 1November 1, 2026: Value only with search or usage data, 30-day attention
What is gaining popularity right now, across models, vendors, harnesses and AI subscriptions, is collected in All AI trends.
Similar models
More models of the same kind as Trinity Large Preview, chosen by model line, developer, size, price and release date.
- Trinity Large ThinkingArcee AI · same developer
- Virtuoso LargeArcee AI · same developer
- Trinity MiniArcee AI · same developer
- Trinity Nano PreviewArcee AI · same developer
- Qwen 3.5 397B A17BAlibaba · similar size
- Nex-N2-ProNex AGI · similar size
- Ornith 1.0 397BDeepReinforce · similar size
- GLM-4.6Z.ai · similar size
All specifications
Architecture
Input and output
Watermarking and provenance
AI watermarking refers to invisible signals embedded in the output. Visible marks and provenance metadata are listed separately. Unknown means there is insufficient evidence for that particular field.
- AI watermarking
- Unknown
- no source
- Watermarking technique
- Not known
- no source
Licence and access
Frequently asked questions about Trinity Large Preview
How large is the context window of Trinity Large Preview?
524,288 tokens. That is how much text Trinity Large Preview reads at once, about 790 A4 pages.
Page count as a rule of thumb from Artificial Analysis
Can I run Trinity Large Preview locally?
Yes, the weights are open. With 398B parameters it needs a lot of memory, more a server than a laptop.
Stated by Arcee AI
- Open weights
- yes
- License
- OpenMDW-1.1
- Parameters
- 398B
- active per token
- 13B
Sources and data date
Show all 10 sources
Benchmark sources at a glance
Which independent evaluators tested this model. This is coverage, not a quality ranking.
No active result from the reviewed external evaluators can currently be attributed to this model. Other source evidence remains available below.
Without vendor claims and historical or disputed results; those stay labelled in the single values.
| Type | Evidence and data date |
|---|---|
| Input | Documented modalities. Retrieved October 27, 2026. |
| Output | Documented modalities. Retrieved October 27, 2026. |
| Research status | Research date October 15, 2026. 2 source URLs checked. This documents the inspected sources, not an exhaustive inventory of every publication. |
| Not published | evaluation-details Note The table does not document prompts, sampling/effort, run dates or units; rows are not treated as reproducible reruns. |
| Unresolved | SimpleQA measurement metadata Note Table 4 reports only a “Score” and does not specify unit, scale, score definition, direction, sample size, confidence interval, evaluated checkpoint, or row-level configuration. |
| Additional source | Hugging Face · Arcee AI Trinity Large Preview model card · Editorial description reviewed October 2, 2026 |
| Additional source | Arcee Trinity Large Preview model-card evaluation table (retrieved October 15, 2026) |
| Additional source | Arcee Trinity Large Technical Report, Table 4 (retrieved October 19, 2026) |
| Additional source | OpenRouter: Popularity index (derived), CC BY 4.0 |
| Additional source | Vercel AI Gateway: Popularity index (derived), CC BY 4.0 |
| Additional source | Ollama: cumulative downloads for the popularity index |
| Additional source | Replicate: cumulative runs for the popularity index |
| Additional source | Google Trending Now: limited feed selection for Germany and the US |
| Additional source | Creative Commons Attribution 4.0 |