Artificial Analysis launched an independent implementation of Handshake AI’s VIALS benchmark on November 6, 2026, focused on scientific image interpretation by vision-language models. Across 161 tasks, Gemini 4 Carbon (high) achieved the highest score at 64%, followed by GPT-6 Astra (max) at 63% and Claude Opus 5.5 (max with fallback) at 61%. GPT-6 Luna (max) completed the same task set for $0.02, compared with $1.77 for Grok 4.7 (xhigh), a cost difference of nearly 99 times. Kimi K3.1 Pro, the leading open-weights model, scored 45%, while fourteen tasks triggered safety refusals after models misread scientific images as potentially sensitive content.
Source: Read the original article
Disclaimer: this content is for information purposes only and is not financial advice. Cryptocurrencies are highly volatile: you may lose all of your capital. Always do your own research. Legal notice
