Similar VIALS scores, opposite strategies: Opus 5.5 makes ~32 tool calls per task vs Astra's ~8
ArtificialAnlys · x · 2026-11-07
Artificial Analysis finds similar VIALS-AA scores can hide very different visual reasoning strategies: Claude Opus 5.5 (max with fallback) averages 32 tool calls per task vs 8 for GPT-6 Astra, while scoring nearly the same (61% vs 63%).
The reason: Opus performs extensive annotation and measurement, averaging 8 Image Measurement calls per task vs 0.4 for Astra, while Image Viewing makes up a larger share of Astra's calls (51% vs 42%). Opus measures its way to answers; Astra looks more and measures less — a striking strategy divergence at equal accuracy.
More from Models
- Inkling 2S matches Dots3 on ARC but costs roughly 3.5x more — teortaxesTex · 2026-11-07
- Open ASR Leaderboard adds private Egypt and Nigeria conversational data for accent breakdown — andimarafioti · 2026-11-07
- Users say Claude Opus 5.5 remains unmatched a month after release — prasenx · 2026-11-07
- Ripple-65B Preview: Apache 2.0 open US model tops TerminalBench 2.1 with $32M R&D — The Decoder · 2026-11-07
- Mirage: A 22MB VLM Runs Detection, OCR and Segmentation in One Pass at 74ms on Phone NPU — Henrie_the_dreamer · 2026-11-07
- 4B Nemotron 3 Nano claimed to beat frontier models on task-specific workflows — eliano · 2026-11-07