# Artificial Analysis finds similar VIALS-AA scores with very different tool use

> Original title: Similar VIALS-AA scores can result from very different approaches to visual reasoning

- Source: Artificial Analysis
- Published: 2026-11-06T17:09:37.000Z
- Bad Influence: https://badinfluence.news/items/nquijbo0a8g19le8iwjg819ii
- Original: https://x.com/ArtificialAnlys/status/2118752679140655572

## Summary

Artificial Analysis reports that Claude Opus 5.5 (max with fallback) and GPT-6 Astra score 61% and 63% on VIALS-AA despite different tool use. Opus averages about 32 tool calls per task, including about 8 Image Measurement calls, while Astra averages about 8 calls with about 0.4 Image Measurement calls. Image Viewing makes up about 51% of Astra's tool calls and about 42% of Opus's.

## Post

Similar VIALS-AA scores can result from very different approaches to visual reasoning

Claude Opus 5.5 (max with fallback) averages ~32 tool calls per task, compared with ~8 for GPT-6 Astra, while their VIALS-AA scores are similar (61% and 63%, respectively). This is because Opus performs extensive annotation and measurement, averaging ~8 Image Measurement tool calls per task, compared to ~0.4 for Astra. Image Viewing accounts for a larger share of Astra’s tool calls at ~51% vs. ~42% for Opus

![](https://badinfluence.news/api/img-proxy?u=https%3A%2F%2Fpbs.twimg.com%2Fmedia%2FHWdSVTsWoAAOs_P.jpg&mode=thumb&exp=1794182400&sig=c310e8bb6855908c)