Expired70% confidenceFactTime unknown
In the Super Mario remake test, Opus nearly reproduced the original game experience after three iterations, while Gemini had inaccurate music and incorrect gameplay after three iterations.
1
Sources
70%
Confidence
Medium-term (~90 days)
Relevance
5/31/2026
First Seen
Valid until: 8/29/2026(expired)
Sources
Related Entities
Related Claims
UnverifiedIn the Pac-Man test, Opus achieved normal game mechanics after two iterations, while after giving Gemini three chances, Pac-Man could not even move.76% similarUnverifiedIn the Kitten Cannon clone game test, Gemini 2.5 Pro generated 636 lines of code on the first attempt with a type error, and after three prompts the game ran successfully.72% similarUnverifiedIn a hands-on coding test, tech blogger Theo found that Gemini 3.5 Flash was the only model among all tested that failed to get the game running properly.71% similarUnverifiedGemini 3 significantly outperforms comparable models on the SWE-bench benchmark in agentic scenarios.71% similarUnverifiedGemini 3.1 Pro is prone to hallucinations and sometimes produces inconsistent output quality.69% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/9556API
curl https://kongchang.com/api/v1/knowledge/claims/9556MCP
get_claim(id=9556)