[KongchangAI]
Unverified60% confidenceBenchmarkExact time

The article claims that Gemini Ultra surpassed human expert-level performance on the MMLU benchmark for the first time and matched or exceeded GPT-4's performance.

2
Sources
60%
Confidence
Long-term
Relevance
8/26/2026
First Seen

Sources

Can You Still Use Google AI Pro Student Discount After Graduation? Terms Breakdown & Strategies

Related Entities

Gemini UltramodelGPT-4modelMMLUconcept

Related Claims

UnverifiedGemini Ultra was the first to surpass human expert-level performance on the MMLU benchmark, achieving a score of 90.0%.76% similar

Cite This Claim

Stable URI
https://kongchang.com/claim/806437
API
curl https://kongchang.com/api/v1/knowledge/claims/806437
MCP
get_claim(id=806437)