Expired50% confidenceFactExact time
The open-source OCR model Seria has 650 million parameters, achieves 83.3% accuracy on the AlmocoBench test, can process 5 pages per second on an RTX 5090, and supports 91 languages as well as table recognition.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
7/2/2026
First Seen
Valid until: 9/30/2026(expired)
Sources
Related Claims
UnverifiedOCR's test benchmark is drawn from 50 open-source repositories and 200 real PRs, covers 10 programming languages, and has 1,505 real defects annotated by more than 80 senior engineers.69% similarUnverifiedIn a hands-on Pokémon RPG game development test, a single AI agent produced approximately 800 lines of code with a feature completeness of approximately 65–70%.66% similarUnverifiedAI application testing requires assertion methods based on semantic equivalence rather than exact string matching, as well as quality evaluation based on statistical distribution (e.g., accuracy must reach above 95% across 1,000 calls).64% similarUnverifiedAlibaba Tongyi Lab released the agent evaluation benchmark Powbench V1.0, which contains 150 real-world tasks and 4,050 test units, with the combination of Quant 3.6 Max Preview and QuantPow achieving first place overall.64% similarUnverifiedMBR解码需要从模型中采样数十到数百个候选译文并对每对候选计算效用得分，计算复杂度为O(n²)62% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/53948API
curl https://kongchang.com/api/v1/knowledge/claims/53948MCP
get_claim(id=53948)