Unverified50% confidenceBenchmarkExact time
On hardcore scientific tasks requiring expert-level knowledge such as TominoBench 3.0/4.0, V4.1 Flash still shows a significant gap compared to top-tier models.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
9/11/2026
First Seen
Valid until: 12/10/2026
Sources
Related Entities
Related Claims
UnverifiedIn the cube unfolding and 3D folding question, both V4.1 Flash and V4 Flash produced blank screens and failed, V4 Pro had a correct unfolding diagram but failed due to incorrect closure logic, and JLM 5.3 Flash performed best.65% similarUnverifiedThis evaluation used six composite questions covering front-end, reasoning, and causality, with V4.1 Flash, V4 Flash, V4 Pro with deep thinking enabled, and JLM 5.3 Flash tested back-to-back on the same questions.64% similarUnverifiedA developer stated that when building a cybersecurity system, locally deployed DeepSeek V4 flash was able to match the best solutions from leading labs.63% similarUnverified不考虑成本时，前端编程能力排名为：参照模型 > Kimi K3 > Grok 4.6 > DeepSeek V4 Pro > DeepSeek V4 Flash62% similarUnverifiedV4.1 Flash took first place on the DeepSWE, CyberDream, and AutomationBench agent benchmarks.62% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/900401API
curl https://kongchang.com/api/v1/knowledge/claims/900401MCP
get_claim(id=900401)