InternVL2.5-40B: benchmark results
Provider: Shanghai AI Lab.
Aggregate Rating 1561 ± 41, rank #1524 of 10916 rated models, from 8 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| FinMTM - Multiple-Choice | 35.2 | FinMTM multiple-choice score (%) on the 1,982 negative-selec | 33.3 |
| FinMTM - Single-Choice | 72.3 | FinMTM single-choice score (%) on the 1,982 single-choice qu | 33.3 |
| FinMTM - Financial Agent | 16.8 | FinMTM financial agent score (0-100) over 1,000 tool-use tas | 23.8 |
| FinMTM - Financial Agent (Fuzzed) | 13.5 | FinMTM financial agent score (0-100) over 1,000 tool-use tas | 23.8 |
| FinMTM - Memory | 26.7 | FinMTM open-ended memory (L4: cross-page entity linking, lon | 19 |
| FinMTM - Calculation | 64.6 | FinMTM open-ended calculation (L2: multi-step numerical calc | 14.3 |
| FinMTM - Comprehension | 66.1 | FinMTM open-ended comprehension (L1: entity recognition and | 14.3 |
| FinMTM - Self-Correction | 36.2 | FinMTM open-ended self-correction (L3: adversarial robustnes | 9.5 |
Interactive version: theaggregate.ai/model?slug=internvl2-5-40b · How It Works · Data refreshed daily, snapshot 2026-11-08.