# DeepSeek V4 Flash 低推理预算下仍近满分

- 来源：Dongxi 东锡 NLP (@dongxi_nlp)
- 发布时间：2026-08-08 05:09
- AIHOT 分数：52
- AIHOT 链接：https://aihot.virxact.com/items/cmsjhkcut0a6jroo592eci7th
- 原文链接：https://x.com/dongxi_nlp/status/2085835844984692910

## AI 摘要

DeepSeek V4 Flash 在 ARC-AGI-1 上以低推理预算即得约 84 分，最高约 89 分，额外推理仅多解 5 题，表明其后训练产生了异常高效的抽象推理。该模型 ARC-AGI-2 得分 61.4%（$0.04/任务），ARC-AGI-1 得分 89.0%（$0.02/任务），树立了性价比帕累托前沿新标准。

## 正文

DeepSeek reaches close to its best score even when given a smaller reasoning budget.

ARC-AGI-1:
Low solves about 84.
Max solves about 89.

Extra reasoning recovers only five additional puzzles.

This suggests DeepSeek's post-training produces unusually efficient abstract reasoning.

### 引用推文

> ARC Prize：DeepSeek V4 Flash from @deepseek_ai on ARC-AGI (Verified): - ARC-AGI-2: 61.4%, $0.04/task - ARC-AGI-1: 89.0%, $0.02/task DeepSeek V4 Flash sets the new standard...