Event ·Multi-source

GPT-6 Astra文档解析基准测试结果

GPT-6 Astra在短文档和中等文档的单次提取任务中表现最佳，短文档准确率达97.2%，中等文档达90.6%。该模型在表格理解方面表现优异，但在图表、布局和语义格式处理上存在不足。每页处理成本为11美分，比经济型解决方案高出10倍。

Current read

GPT-6 Astra在文档提取任务上创SOTA，但成本高昂且长文档处理效果差。

11 sources73° AI heatLast updated 2026/9/5 15:25:59

Evidence

11 sources

Primary sourceJerry Liu

2026/9/5 15:25:59View source

Related source 1The Rundown AI

2026/9/3 18:07:36View source

Related source 2Decoder

2026/9/3 19:25:40View source

Related source 3OpenAI

2026/9/3 19:32:14View source

Related source 4@OpenAIDevs

2026/9/3 19:33:20View source

Related source 5ChatGPT

2026/9/3 19:39:13View source

Related source 6Cognition

2026/9/3 19:40:15View source

Related source 7Greg Brockman

2026/9/3 19:44:36View source

Related source 8宝玉

2026/9/3 19:48:56View source

Related source 9Sam Altman

2026/9/3 19:49:10View source

Related source 10elvis

2026/9/3 19:55:42View source

Conflict check

Dedup data only shows that these sources discuss the same event, not that they agree. There is no structured supporting or refuting evidence yet, so no conflict is inferred automatically.

Open the story