Ling-3.0-flash-fin: benchmark results
Ant Group's open-weight Ling-3.0-flash-Fin (August 2026), the first finance-specialized model in the Ling family, extending Ling-3.0-flash with continued training on financial data, 124B total and 5.1B active parameters. Provider: InclusionAI. Released 2026-08-28. Access: Open.
Aggregate Rating 1754 ± 23, rank #109 of 1635 rated models, from 83 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| LLM Stats (Tau3 Banking) | 41 | Score (%) | 100 |
| FrontierFinance - Company Research | 47.79 | Rubric qualification rate (%, macro average over queries, th | 91.7 |
| AA-Omniscience Hallucination Rate | 39.51 | Hallucination Rate (%) | 87.3 |
| LLM Stats Score | 42.68 | LLM Stats Score (conservative rating) | 85.8 |
| FrontierFinance - Coverage & Catalyst Monitoring | 48.24 | Rubric qualification rate (%, macro average over queries, th | 83.3 |
| FrontierFinance - Earnings & Events | 61.33 | Rubric qualification rate (%, macro average over queries, th | 83.3 |
| Tau3 Banking | 38.56 | Success Rate (%) | 82.5 |
| AA-Omniscience Index - Software Engineering (SWE) - Dart | 14 | Omniscience Index | 79.7 |
| FinFIRST - Source Verification | 82.45 | Weighted rubric score (%; the 2,085 rubric points for using | 78.6 |
| Tinybird AI SQL Benchmark - Success Rate | 100 | Questions answered with a valid query within 3 retries (%) | 71.7 |
| Artificial Analysis Intelligence Index | 22.6 | Intelligence Index | 70.9 |
| AA Humanity's Last Exam | 22.61 | Accuracy (%) | 70.2 |
Interactive version: theaggregate.ai/model?slug=ling-3-0-flash-fin · How It Works · Data refreshed daily, snapshot 2026-10-15.