Anthropic models use more tools as they improve, OpenAI models fewer: VIALS-AA data
ArtificialAnlys · x · 2026-11-10
Artificial Analysis built VIALS-AA from Handshake AI's VIALS life-sciences visual reasoning benchmark (gel blots, plasmid maps, phylogenies). Claude Fable 5.2 (max) leads at 67% pass^3, with Gemini 4 Carbon at 64% and GPT-6 Astra at 63%. Key split: Anthropic's tool calls rise with capability, from 21 per task (Haiku 5.5) to 37 (Fable 5.2), driven mostly by image viewing (+30% vs Opus, code execution only +5%). OpenAI models go the opposite way — GPT-6 Astra, GPT-6.1 Sol, GPT-6.1 Terra and GPT-6 Luna all make fewer tool calls than Haiku, with the highest-scoring Astra making the fewest.
More from Models
- Claude Fable 5.2 scores 64.9% on RoboDojo robot control, dominating rivals — burny_tech · 2026-11-10
- Claude Fable 5.2 video demo wows on pacing, but AI-sounding music persists — jiayuan_jy · 2026-11-10
- Gemini Notebooks keeps pulling context from old conversations, user reports — AnonymousDork929 · 2026-11-10
- User complains model takes four minutes to compare two 16-token completions — cephaloform · 2026-11-10
- DavidAU releases Qwen3.8-40B-Ultimate-Heretic-Uncensored-Turbo-Brilliance-50x-AGI-NEO-GGUF merge — mp3m4k3r · 2026-11-10
- Uncensored Gemma 4 31B 'heretic' build hits Hugging Face, 11.9GB quantized — alexcovo_eth · 2026-11-10