Language modelOpen weights
Ling-3.0-flash-Fin
InclusionAI
- Released
- -
- Data date
- October 3, 2026
Ling-3.0-flash-Fin is InclusionAI’s first finance-specialized model in the Ling series, according to the company. It builds on Ling 3.0 Flash with continued training on financial data. Its focus includes reconciling reporting periods and figures, valuation models, and spreadsheet formulas.
The MoE checkpoint activates 5.1 of its 124 billion parameters per token and handles 256K context tokens. InclusionAI releases it under MIT, with access locally and through OpenRouter.
Specifications and access
| Specification | Value and source |
|---|---|
| Model ID | inclusionAI/Ling-3.0-flash-FinSource |
| Model class | Finance-specialized language modelSource |
| Parameters | 124 billion total, 5.1 billion activeSource |
| Context window | 256K tokens according to the model cardSource |
| Architecture | Ling-3.0-flash architecture with a mixture of expertsSource |
| Input | Text and financial documentsSource |
| Access | OpenRouter and local deploymentSource |
| License | MITSource |
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.
FinFIRST (Show measurement, test conditions, and source)
- Source value
- 52.85
- Score
- 52.85
- Metric
- FinFIRST reported score
- Unit
- No unit provided
- Category
- source-specific
- Direction
- Higher is better
- Source type
- vendor-reported
- Evaluator
- Provider model card
- Status
- active
- Retrieved at
- 2026-10-04
- Methodology
- Vendor-reported result. Official image-only table, visually inspected. ReAct/Web Search/Visit/Python and Claude Code spreadsheet harnesses; highest reasoning tier, mixed external and own evaluations.
- Context
- Source-specific observation; it is not a shared comparison cohort. The source table does not state a unit or scale for this numeric score.
FinSearchComp Verified (Show measurement, test conditions, and source)
- Source value
- 77.04
- Score
- 77.04
- Metric
- FinSearchComp Verified reported score
- Unit
- No unit provided
- Category
- source-specific
- Direction
- Higher is better
- Source type
- vendor-reported
- Evaluator
- Provider model card
- Status
- active
- Retrieved at
- 2026-10-04
- Methodology
- Vendor-reported result. Official image-only table, visually inspected. ReAct/Web Search/Visit/Python and Claude Code spreadsheet harnesses; highest reasoning tier, mixed external and own evaluations.
- Context
- Source-specific observation; it is not a shared comparison cohort. The source table does not state a unit or scale for this numeric score.
FinCRAFT (Show measurement, test conditions, and source)
- Source value
- 51.74
- Score
- 51.74
- Metric
- FinCRAFT reported score
- Unit
- No unit provided
- Category
- source-specific
- Direction
- Higher is better
- Source type
- vendor-reported
- Evaluator
- Provider model card
- Status
- active
- Retrieved at
- 2026-10-04
- Methodology
- Vendor-reported result. Official image-only table, visually inspected. ReAct/Web Search/Visit/Python and Claude Code spreadsheet harnesses; highest reasoning tier, mixed external and own evaluations.
- Context
- Source-specific observation; it is not a shared comparison cohort. The source table does not state a unit or scale for this numeric score.
Finance Agent v1.1 (Show measurement, test conditions, and source)
- Source value
- 69.19
- Score
- 69.19
- Metric
- Finance Agent v1.1 reported score
- Unit
- No unit provided
- Category
- source-specific
- Direction
- Higher is better
- Source type
- vendor-reported
- Evaluator
- Provider model card
- Status
- active
- Retrieved at
- 2026-10-04
- Methodology
- Vendor-reported result. Official image-only table, visually inspected. ReAct/Web Search/Visit/Python and Claude Code spreadsheet harnesses; highest reasoning tier, mixed external and own evaluations.
- Context
- Source-specific observation; it is not a shared comparison cohort. The source table does not state a unit or scale for this numeric score.
Finance Agent v2 (Show measurement, test conditions, and source)
- Source value
- 59.81
- Score
- 59.81
- Metric
- Finance Agent v2 reported score
- Unit
- No unit provided
- Category
- source-specific
- Direction
- Higher is better
- Source type
- vendor-reported
- Evaluator
- Provider model card
- Status
- active
- Retrieved at
- 2026-10-04
- Methodology
- Vendor-reported result. Official image-only table, visually inspected. ReAct/Web Search/Visit/Python and Claude Code spreadsheet harnesses; highest reasoning tier, mixed external and own evaluations.
- Context
- Source-specific observation; it is not a shared comparison cohort. The source table does not state a unit or scale for this numeric score.
APEX-Agents (Show measurement, test conditions, and source)
- Source value
- 29.17
- Score
- 29.17
- Metric
- APEX-Agents reported score
- Unit
- No unit provided
- Category
- source-specific
- Direction
- Higher is better
- Source type
- vendor-reported
- Evaluator
- Provider model card
- Status
- active
- Retrieved at
- 2026-10-04
- Methodology
- Vendor-reported result. Official image-only table, visually inspected. ReAct/Web Search/Visit/Python and Claude Code spreadsheet harnesses; highest reasoning tier, mixed external and own evaluations.
- Context
- Source-specific observation; it is not a shared comparison cohort. The source table does not state a unit or scale for this numeric score.
SpreadsheetBench V1 (Show measurement, test conditions, and source)
- Source value
- 86.5
- Score
- 86.5
- Metric
- SpreadsheetBench V1 reported score
- Unit
- No unit provided
- Category
- source-specific
- Direction
- Higher is better
- Source type
- vendor-reported
- Evaluator
- Provider model card
- Status
- active
- Retrieved at
- 2026-10-04
- Methodology
- Vendor-reported result. Official image-only table, visually inspected. ReAct/Web Search/Visit/Python and Claude Code spreadsheet harnesses; highest reasoning tier, mixed external and own evaluations.
- Context
- Source-specific observation; it is not a shared comparison cohort. The source table does not state a unit or scale for this numeric score.
SpreadsheetBench V2 (Show measurement, test conditions, and source)
- Source value
- 21.81
- Score
- 21.81
- Metric
- SpreadsheetBench V2 reported score
- Unit
- No unit provided
- Category
- source-specific
- Direction
- Higher is better
- Source type
- vendor-reported
- Evaluator
- Provider model card
- Status
- active
- Retrieved at
- 2026-10-04
- Methodology
- Vendor-reported result. Official image-only table, visually inspected. ReAct/Web Search/Visit/Python and Claude Code spreadsheet harnesses; highest reasoning tier, mixed external and own evaluations.
- Context
- Source-specific observation; it is not a shared comparison cohort. The source table does not state a unit or scale for this numeric score.
tau3-Banking (Show measurement, test conditions, and source)
- Source value
- 41
- Score
- 41
- Metric
- tau3-Banking reported score
- Unit
- No unit provided
- Category
- source-specific
- Direction
- Higher is better
- Source type
- vendor-reported
- Evaluator
- Provider model card
- Status
- active
- Retrieved at
- 2026-10-04
- Methodology
- Vendor-reported result. Official image-only table, visually inspected. ReAct/Web Search/Visit/Python and Claude Code spreadsheet harnesses; highest reasoning tier, mixed external and own evaluations.
- Context
- Source-specific observation; it is not a shared comparison cohort. The source table does not state a unit or scale for this numeric score.
Sources and data date
Every statement links to its underlying documentation or leaderboard.
| Type | Evidence and data date |
|---|---|
| Research status | Research date October 4, 2026. 1 source URLs checked. This documents the inspected sources, not an exhaustive inventory of every publication. |
| Additional source | Ling-3.0-flash-Fin model card (retrieved October 3, 2026; October 4, 2026) · Ling-3.0-flash-Fin model card · Editorial description reviewed October 4, 2026 |
| Additional source | Ling-3.0-flash-VL model card (retrieved October 3, 2026) · Ling-3.0-flash-VL model card · Editorial description reviewed October 4, 2026 |
| Additional source | Ling-3.0-flash base model card (retrieved October 3, 2026) · Ling-3.0-flash base model card · Editorial description reviewed October 4, 2026 |