Unverified50% confidenceEventExact time
Bilibili creator "Da Fei Yu" released a video self-described as "the least objective model review of the year", which used no benchmarks or parameters and rated multiple large models solely by user experience and emotional value.
1
Sources
50%
Confidence
Short-term (~14 days)
Relevance
10/24/2026
First Seen
Valid until: 11/7/2026
Sources
Related Entities
Related Claims
UnverifiedPublic leaderboards reflect which model has received the most upvotes, not which model simultaneously satisfies multiple specific constraints in a single run.62% similarUnverifiedThe video author believes that raising a model's floor (worst-case performance) is more valuable than raising its ceiling (best-case performance)61% similarUnverifiedThe standard evaluation metrics for AI video generation models mainly include FID, FVD, and CLIP Score, but these metrics have difficulty capturing fine-grained errors that require domain knowledge and physical common sense to identify.61% similarUnverifiedAccording to analyses by multiple science popularization channels on Bilibili and YouTube, the sources are basically consistent on the figure that "only 9% of people can correctly predict," but do not disclose in detail the specific experimental sample sizes and statistical methods.57% similarUnverifiedWhen comparing models horizontally on making short films, it is necessary to unify the criteria such as reasoning effort, whether caching is included, and whether tool-call tokens are included; simply comparing time and token usage is unfair.56% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1092369API
curl https://kongchang.com/api/v1/knowledge/claims/1092369MCP
get_claim(id=1092369)