Language modelOpen weights
Apodex 1.1 Mini
Apodex
- Released
- September 1, 2026
- Data date
- October 7, 2026
- Input
- Text and images
- Output
- Text
Apodex 1.1 Mini is the open text-and-image variant Apodex released on September 1, 2026. The provider announces 35 billion parameters, while Hugging Face lists 36 billion. The weights use Apache 2.0.
Specifications and access
| Specification | Value and source |
|---|---|
| Provider model ID | apodex/Apodex-1.1-miniSource |
| Access | Open weights for local deploymentSource |
| Input modalities | Text and imagesSource |
| Output modalities | TextSource |
| License | Apache 2.0Source |
| Announced parameters | 35 billionSource |
| Weight-metadata parameters | Hugging Face lists 36 billion; differs from the 35B announcementSource |
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.
FrontierFinance (Show measurement, test conditions, and source)
- Source value
- 40
- Score
- 40
- Metric
- rubric qualification rate, macro-averaged over 220 questions; unanswered tasks score zero
- Unit
- %
- Category
- agentic
- Direction
- Higher is better
- Sample
- 220
- Source type
- vendor-reported
- Evaluator
- Apodex 1.1 technical report
- Status
- active
- Retrieved at
- 2026-10-07
- Methodology
- Vendor-reported result. Apodex 1.1 Mini model with the paper's ReAct harness. This is a model-plus-harness result, not an isolated base-checkpoint score. The paper names each benchmark's task metric and sample size where published.
- Context
- Provider-reported system results. ReAct and Agent Team are separate harness configurations; they must not be presented as checkpoint-only measurements or merged with flagship Apodex 1.1 results.
FrontierFinance (Show measurement, test conditions, and source)
- Source value
- 50.2
- Score
- 50.2
- Metric
- rubric qualification rate, macro-averaged over 220 questions; unanswered tasks score zero
- Unit
- %
- Category
- agentic
- Direction
- Higher is better
- Sample
- 220
- Source type
- vendor-reported
- Evaluator
- Apodex 1.1 technical report
- Status
- active
- Retrieved at
- 2026-10-07
- Methodology
- Vendor-reported result. Apodex 1.1 Mini with the paper's Agent Team multi-agent harness. This is a model-plus-harness result, not an isolated base-checkpoint score. The paper names each benchmark's task metric and sample size where published.
- Context
- Provider-reported system results. ReAct and Agent Team are separate harness configurations; they must not be presented as checkpoint-only measurements or merged with flagship Apodex 1.1 results.
FrontierScience-Research (Show measurement, test conditions, and source)
- Source value
- 45
- Score
- 45
- Metric
- fraction of tasks passing at least 7/10 rubric points
- Unit
- %
- Category
- agentic
- Direction
- Higher is better
- Sample
- 60
- Source type
- vendor-reported
- Evaluator
- Apodex 1.1 technical report
- Status
- active
- Retrieved at
- 2026-10-07
- Methodology
- Vendor-reported result. Apodex 1.1 Mini model with the paper's ReAct harness. This is a model-plus-harness result, not an isolated base-checkpoint score. The paper names each benchmark's task metric and sample size where published.
- Context
- Provider-reported system results. ReAct and Agent Team are separate harness configurations; they must not be presented as checkpoint-only measurements or merged with flagship Apodex 1.1 results.
FrontierScience-Research (Show measurement, test conditions, and source)
- Source value
- 51.7
- Score
- 51.7
- Metric
- fraction of tasks passing at least 7/10 rubric points
- Unit
- %
- Category
- agentic
- Direction
- Higher is better
- Sample
- 60
- Source type
- vendor-reported
- Evaluator
- Apodex 1.1 technical report
- Status
- active
- Retrieved at
- 2026-10-07
- Methodology
- Vendor-reported result. Apodex 1.1 Mini with the paper's Agent Team multi-agent harness. This is a model-plus-harness result, not an isolated base-checkpoint score. The paper names each benchmark's task metric and sample size where published.
- Context
- Provider-reported system results. ReAct and Agent Team are separate harness configurations; they must not be presented as checkpoint-only measurements or merged with flagship Apodex 1.1 results.
APEX-Agents (Show measurement, test conditions, and source)
- Source value
- 24.2
- Score
- 24.2
- Metric
- reported APEX-Agents score; paper does not state a unit in the mini-model table
- Unit
- No unit provided
- Category
- agentic
- Direction
- Higher is better
- Sample
- 480
- Source type
- vendor-reported
- Evaluator
- Apodex 1.1 technical report
- Status
- active
- Retrieved at
- 2026-10-07
- Methodology
- Vendor-reported result. Apodex 1.1 Mini model with the paper's ReAct harness. This is a model-plus-harness result, not an isolated base-checkpoint score. The paper names each benchmark's task metric and sample size where published.
- Context
- Provider-reported system results. ReAct and Agent Team are separate harness configurations; they must not be presented as checkpoint-only measurements or merged with flagship Apodex 1.1 results.
APEX-Agents (Show measurement, test conditions, and source)
- Source value
- 27.7
- Score
- 27.7
- Metric
- reported APEX-Agents score; paper does not state a unit in the mini-model table
- Unit
- No unit provided
- Category
- agentic
- Direction
- Higher is better
- Sample
- 480
- Source type
- vendor-reported
- Evaluator
- Apodex 1.1 technical report
- Status
- active
- Retrieved at
- 2026-10-07
- Methodology
- Vendor-reported result. Apodex 1.1 Mini with the paper's Agent Team multi-agent harness. This is a model-plus-harness result, not an isolated base-checkpoint score. The paper names each benchmark's task metric and sample size where published.
- Context
- Provider-reported system results. ReAct and Agent Team are separate harness configurations; they must not be presented as checkpoint-only measurements or merged with flagship Apodex 1.1 results.
Sources and data date
| Type | Evidence and data date |
|---|---|
| Input | Documented modalities. Retrieved October 7, 2026. |
| Output | Documented modalities. Retrieved October 7, 2026. |
| Research status | Research date October 7, 2026. 4 source URLs checked. This documents the inspected sources, not an exhaustive inventory of every publication. |
| Not found in the inspected sources | independent-mini-results Original wording No independent evaluation with published numbers for the exact Apodex 1.1 Mini was verified in this pass. The values come from two explicitly separate provider harnesses. |
| Additional source | Apodex 1.1 Mini model card (retrieved October 7, 2026) · Apodex 1.1 Mini model card · Editorial description reviewed October 7, 2026 |
| Additional source | Apodex 1.1 technical report (retrieved October 7, 2026) |
| Additional source | Model metadata (retrieved October 7, 2026) |