# GPT-6 Astra 基准测试对比 GPT-5.6 Sol 与 Claude Fable 5.1，ARC-AGI-3 从 7.8% 升至 98.6%

- 来源：Chubby♨️ (@kimmonismus)
- 发布时间：2026-09-04 02:45
- AIHOT 分数：24
- AIHOT 链接：https://aihot.virxact.com/items/cmtlwgygo0pnprow5t9xcoyy4
- 原文链接：https://x.com/kimmonismus/status/2095583875611140185

## AI 摘要

作者转发一张完整基准对比表，称 OpenAI 新发布的 GPT-6 Astra 在多项评测上明显领先。ARC-AGI-3 从 GPT-5.6 Sol 的 7.8% 升至 98.6%，FrontierMath Tier 4 (v2) 为 97.6%，SRE-Bench 为 99.2%，ExploitBench 为 100.0%；作者称其压过了新发布的 Claude Fable 5.1。

## 正文

You can quote me on this: OpenAI crushed Anthropic’s IPO.

They put the newly released, state-of-the-art Fable 5.1 to shame. And they knew exactly what they were doing.

Just look at the numbers. Its not even close.

### 引用推文

> Chubby♨️：Full benchmarks. Absolutely insane. ARGI-AGI 3, from 7.8% to 98.6%