Unverified50% confidenceBenchmarkExact time
Gemini 4's success rate in tasks simulating humans operating a computer is less than 5%, ranking second to last among 8 models.
1
Sources
50%
Confidence
Short-term (~14 days)
Relevance
10/14/2026
First Seen
Valid until: 10/28/2026
Sources
Related Entities
Related Claims
UnverifiedAccording to Bloomberg, Google's internal employees are not satisfied with Gemini 4's actual performance, believing the model gets stuck on certain programming tasks.75% similarUnverifiedBloomberg reported that while Gemini 4 performs well on industry benchmarks, it performs less well when employees use it in their actual work.74% similarUnverifiedAccording to Google's results on the GraySwan benchmark, Gemini 4 Argon has the lowest attack success rate among all tested models.73% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1030256API
curl https://kongchang.com/api/v1/knowledge/claims/1030256MCP
get_claim(id=1030256)