Trinity family
Open-weight reasoning models for local experimentation and enterprise deployment; Trinity Mini is 26B total with 3B active parameters.
Score calculation
(2 + 2 + 6 + 1 + 7) ÷ 5 = 3.6
5 of six dimensions rated. Unrated dimensions are excluded from the average.
Full scoring method ↗Setup details
Arcee Trinity family, including published Nano, Mini, and Large checkpoints; license and deployment requirements vary by release.
What fits your needs
Arcee Trinity family, including published Nano, Mini, and Large checkpoints; license and deployment requirements vary by release.
Good fit
Open-weight reasoning models for local experimentation and enterprise deployment; Trinity Mini is 26B total with 3B active parameters.
Tradeoffs to expect
Tradeoffs
Trinity Mini model card lists OpenMDW-1.1; review that model-specific license rather than assuming an OSI-approved or unrestricted license. The model weights and technical details are available, but OpenMDW-1.1 is a model-specific license, not Apache or MIT. Public checkpoints do not by themselves provide the training-data source or rights trail.
How to switch
Local setup
Trinity Mini has 26B total parameters (3B active); plan memory for total weights and context, not only active parameters. Larger variants need server-class hardware.
Practical questions
Weights and license
Fetch the exact checkpoint revision and read its own license; sibling and distilled variants may inherit different terms.
Hardware
Trinity Mini has 26B total parameters (3B active); plan memory for total weights and context, not only active parameters. Larger variants need server-class hardware.
Training data
Treat the weight license and public web/data availability separately from evidence of permission for the training data.
This is a practices and disclosure score for the named release, not a request-level energy or emissions estimate.
See the evidence
The Trinity Mini card reports a 512-H200 training cluster, but does not measure electricity, emissions, or water for the run. The rubric rewards documented reporting and controls; it does not infer high consumption from missing measurements.
1 + 1 reporting + 0 electricity + 0 controls = 2/10. This grades documented practices and transparency.
No comparable per-request electricity, emissions, water, regional serving mix, or independently reviewed model footprint is established in the cited sources.
- Reporting1/3
- The Trinity Mini card reports a 512-H200 training cluster, but does not measure electricity, emissions, or water for the run.
- Electricity sourcing0/3
- No model-specific clean-electricity sourcing or hourly power evidence is published in the reviewed sources; zero means unestablished, not dirty electricity.
- Resource controls0/3
- No model-specific documented control or measured physical resource reduction is established by the reviewed release materials. Public weights or smaller variants alone do not prove reduced resource use.
Water disclosure
Water reporting has not been assessed for this choice.
The release provides some training-process information, but not a complete permission ledger for inherited and later training data.
See the evidence
Arcee describes a Datology data-curation partnership and a broad token scale, but the reviewed material does not publish an itemized source manifest or training-data permission evidence. A public weight license governs use of those weights; it is not proof that every training source was permitted.
1 base + 1 source transparency + 0 documented permission + 0 creator controls = 2/10. Weight licensing is not treated as training data permission.
Complete source lists, source-specific rights, inherited-base provenance, synthetic-data generation permissions, and creator removal or retraining remedies remain unestablished.
- Source transparency1/3
- Arcee describes a Datology data-curation partnership and a broad token scale, but the reviewed material does not publish an itemized source manifest or training-data permission evidence.
- Documented permission0/4
- A model-weight license governs the released checkpoint, not automatically the training data. The reviewed release doesn’t confirm affirmative rights for all inherited and post-training material.
- Creator controls0/2
- The reviewed model documentation doesn’t show creator permission covering all training data, opt-out, removal, or retraining remedy for this model family.
Privacy depends on where these released weights are run; the model family has no single prompt-retention or training route.
See the evidence
The official release makes weights available for separate deployments, but it does not define one inference host or retention policy for every user. Local and hosted operators may handle prompts differently, so a single numerical privacy grade would misstate this model-level entry.
The user-selected host, region, telemetry, logging, extensions, and provider training terms are outside the checkpoint release and require a route-specific review.
The model weights and technical details are available, but OpenMDW-1.1 is a model-specific license, not Apache or MIT. Public checkpoints do not by themselves provide the training-data source or rights trail.
See the evidence
The model weights and technical details are available, but OpenMDW-1.1 is a model-specific license, not Apache or MIT. Public checkpoints do not by themselves provide the training-data source or rights trail. This openness rating concerns artifacts and reproducibility only; it does not assess whether the original training data was authorized.
Trinity Mini model card lists OpenMDW-1.1; review that model-specific license rather than assuming an OSI-approved or unrestricted license. Model weights and documentation are available, but limitations described above reduce source-level reproducibility.
The checked release and license apply to the named checkpoints; sibling variants, future versions, and the underlying training data may have different terms or disclosures.
Arcee reports that its team built Trinity and will expand; reviewed primary sources do not describe workforce conditions or worker channels.
See the evidence
Arcee’s company update says the Trinity team built its 2025 model lineup and that the Series B will support hiring and growth. It does not set out pay, working conditions, supplier safeguards, or grievance processes for employees and model contractors, so this remains a disclosure gap.
1 base + 0 documented conditions + 0 voice/remedy + 0 assurance = 1/10. Team and funding statements don’t confirm worker protections.
No worker pay, location, conditions, supplier coverage, voice or remedy system, independent audit, or follow-through outcomes were established in the reviewed official company sources.
- Pay & conditions0/3
- Arcee discusses its team and hiring plans, but the reviewed company sources do not publish pay or working-condition standards for staff and contractors.
- Worker voice & remedy0/3
- No employee or supplier grievance channel, non-retaliation commitment, or remedy mechanism was established in the reviewed sources.
- Evidence & follow-through0/3
- No independent worker-level audit or labor remediation outcome was established in the reviewed sources.
Arcee publicly describes a civilian-science collaboration with DOE; no Arcee-Anduril or defense deployment link was established in reviewed official company materials.
See the evidence
Arcee and DOE describe Genesis-Science-1 as an open model for scientific research with national laboratories shaping research tasks. That is a public civilian-government collaboration, not evidence of military use. Trinity’s OpenMDW license imposes no field-of-use restriction; no separate defense safeguard was found.
7 base + 0 safeguard adjustment = 7/10 (2 of 9 safeguard points; every 3 adds 1). No military relationship was established in the cited sources. This ranks above known defense ties without claiming a verified clean record.
No military customer, Anduril relationship, defense contract, Granite-like model deployment, or military-use enforcement process is established in the reviewed official sources.
- Use restrictions0/4
- OpenMDW-1.1 grants unrestricted use without a field-of-use restriction; that is a license scope finding, not proof of military deployment.
- Enforcement & oversight0/3
- No monitoring or enforcement for military use of downloaded Trinity weights was established.
- Relationship disclosure2/2
- Arcee publicly describes its DOE and national-laboratory Genesis collaboration; the source scopes it to scientific research.
Arcee publicly collaborates with DOE and national laboratories on a scientific model. Reviewed company sources did not establish a defense relationship.
Military rubric ↗Arcee AI is not named in the two official BNC campaign lists checked. This is a limited list check, not BDS approval.
See the evidence
The official BNC consumer boycott/pressure guide and No Tech campaign page do not name Arcee AI among their current targets. BDS uses selective campaigns, so absence from these lists is neutral and should not be read as a clean record or approval.
This check covers only named targets in the two cited BNC campaigns; it does not cover all BDS campaigns, local groups, investments, suppliers, or downstream hosts. Lists and ownership may change.
All sources
- Official model card or model release repository ↗
- Official model card or model release repository ↗
- Official model card or model release repository ↗
- BNC official campaign target list ↗
- BNC official campaign target list ↗
- Genesis-Science-1 ↗
- Arcee AI Raises Series B to Build the Future of American Open Models ↗
- OpenMDW FAQ ↗