AITOPAI热报
Models

V4.1 Flash-1017 在 ARC-AGI-2 上略超 Gemini 3.1 Pro Preview

Teortaxes@teortaxesTex
Post
View the post
Why we picked it

V4.1 Flash-1017 在 ARC-AGI-2 上小胜 Gemini 3.1 Pro Preview，想看推理模型基准对比的可以看看这组数据。

某实验室发布了 V4.1 Flash-1017 版本，在 ARC-AGI-2 基准上相比前作有温和提升。该成绩略微超过了 Gemini 3.1 Pro Preview，而两者发布时间相隔 8 个月。作者同时希望官方把 token 消耗作为网页端的可选坐标轴展示。

Original · Teortaxes

🫡 This is my actual role as an “influencer” here I guess Influencing the quality of discourse upwards. That said, I still want to see tokens as an x-axis option *on the website* V4.1 Flash-1017 modestly improves on ARC-AGI-2, slightly exceeding… Gemini 3.1 Pro Preview. 8 months https://t.co/wEUOxSbdBL

#V4.1 Flash#Gemini 3.1 Pro#ARC-AGI-2#推理模型#基准测试

Keep reading

Recent stories with the same tag or section
ModelsUS vs ChinaOfficialSelected21:01
DeepSeek V4.1 Flash tested on game effect prompts

The practical comparison suggests DeepSeek's lightweight tier can handle production-style creative coding tasks that previously demanded Claude's top model. Developers building effect or animation workflows on a budget can treat V4.1 Flash as a workable fallback. Testing against your own Opus prompts is the quickest way to gauge the gap.

#DeepSeek#Claude#coding assistants#open-source models
官方一手歸藏(guizang.ai)@op7418Source