Expired50% confidenceBenchmarkExact time
In an experiment reproducing a physics-related paper, GPT-6 Astra performed best, Opus 5.4 was slightly behind, and DeepSeek has room for improvement on tasks such as rendering.
1
Sources
50%
Confidence
Short-term (~14 days)
Relevance
9/19/2026
First Seen
This claim has expired. It is kept for the record and no longer describes the current state.
Valid until: 10/3/2026(expired)
Sources
Related Entities
Related Claims
UnverifiedA new search agent has demonstrated performance surpassing GPT-6 Astra across multiple benchmarks.77% similarUnverifiedGPT-6 Astra falls short of dedicated tools in professional detail control for 3D modeling, including topology quality, UV layout, and material accuracy.77% similarUnverifiedBilibili uploader Elstra conducted continuous testing of GPT-6 Astra's automatic modeling capability, selecting a building case of larger scale and more complex structural detail.76% similarUnverifiedCommunity comparisons of GPT-6 Astra's initial launch outputs with recently generated results have found that the newer version shows decreased realism and clarity when generating complex 3D scenes.75% similarUnverifiedGPT-6 Astra reached a saturation level of 95.9% in 3D object creation capability tests, more than 10 percentage points higher than competing products.75% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/934798API
curl https://kongchang.com/api/v1/knowledge/claims/934798MCP
get_claim(id=934798)