GPT-6 scores 0.53 on 250-question 3D spatial reasoning test; Claude fails
Overview
AI Era (新智元) reports that GPT-6 Astra scored 0.53 on a new benchmark of 250 3D spatial reasoning questions, and that Claude also failed the test.
The report does not give the scoring scale or how "failed" was defined for Claude. The source excerpt was truncated, so the test methodology, the scoring details and the Claude results are not available.
Written by AI from the articles below · updated Oct 30, 9:21 PM ET
Check the sources:
Article timeline
The articles in this story. Times are ET.
- AI EraNewsGPT-6 scores 0.53 out of 250 on a 3D spatial reasoning test
AIA new benchmark of 250 3D spatial reasoning questions found GPT-6 Astra scored 0.53 points, and Claude failed the test as well. The source excerpt was truncated, so the full test methodology and scoring details are not available.
Heat Trend
Not enough continuous observations to show a trend yet.