Unverified50% confidenceEventExact time
@lnkiai's experiment shows: faced with the same "make a 3D avatar" task, Codex (GPT-6.1 Sol) chose to model from scratch in Blender, while Claude Code (Opus 5.5) proposed finding distributed models and parts and then assembling and machining them.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
10/13/2026
First Seen
Valid until: 1/11/2027
Sources
Related Entities
Related Claims
VerifiedWillison pasted the generated image into Codex running GPT-6 Astra (high) and instructed it to create a 3D model using the local Blender skill.77% similarUnverifiedThe Blender sofa model generated by GPT-6 Sol from an image outperforms Opus 5.5 in handling curved surfaces.72% similarUnverifiedSimon asked GPT-5.6-Sol for tool suggestions, and the model proactively built a prototype of the tool.67% similarUnverifiedGPT-6 Sol exhibited lid clipping and component interpenetration issues in the 3D contact lens case task, with Opus ultimately scoring 10 points and Sol scoring 6 points.67% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1017633API
curl https://kongchang.com/api/v1/knowledge/claims/1017633MCP
get_claim(id=1017633)