Published at 11/12/2026, 1:10:59 AM · 104 views
📰 Artificial Analysis Reruns VIALS-AA Without Tools: Tool Use Nearly Doubles Average Scores From ~25% to ~47%, and Claude Haiku 5.5 With Tools Beats Fable 5.2…
Independent evaluator Artificial Analysis reran VIALS-AA, its life-sciences visual reasoning benchmark, in a one-shot format with one request and no tools. Tool use lifted the cross-model average from about 25% to about 47%.
💬 What's your take? Join the discussion in the comments.
Read the original Chinese version
📰 Artificial Analysis 拿掉工具重跑 VIALS-AA：工具让平均分从约 25% 翻倍到约 47%，带工具的 Claude Haiku 5.5 胜过不带工具的 Fable 5.2，成本仅约十分之一
独立评测机构 Artificial Analysis 把生命科学视觉推理基准 VIALS-AA 改成“一次请求、不给工具”的单发模式重跑。结果显示，工具让各模型平均分从约 25% 升到约 47%。
💬 你怎么看？欢迎在评论区讨论。
#Artificial Analysis #VIALS-AA #Benchmark
0 Comments