Unverified50% confidenceBenchmarkExact time
The model was evaluated in real time on the IOI 2026 problem set, adhering to official time limits, internet access, and submission constraints, ultimately achieving a score of 535.4 out of 600.
1
Sources
50%
Confidence
Long-term
Relevance
9/29/2026
First Seen
Sources
Related Entities
Related Claims
VerifiedThe AI executive order issued by the Biden administration in 2023 requires that large models exceeding a computational threshold of approximately 10²⁶ floating-point operations report their safety testing results to the government.71% similarUnverifiedAcemoglu published research in 2024 indicating that the proportion of tasks current AI can fully automate is approximately 5%.66% similarUnverified欧盟数据保护委员会（EDPB）于2023年专项调查多家AI服务商的训练数据采集合规性66% similarUnverified到2026年，评判AI编程工具的核心指标将变为四个系统级能力：能不能分工、能不能隔离、能不能审计、能不能自己跑到交付66% similarUnverifiedIn early 2023, accurately analyzing a 50-page PDF document with AI was still a technical challenge; by 2024, with the maturation of long-context window models and multimodal capabilities, it had become a basic feature.66% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/958397API
curl https://kongchang.com/api/v1/knowledge/claims/958397MCP
get_claim(id=958397)