Expired50% confidenceFactTime unknown
In the OmniDocBench V1.5 benchmark, DeepSeek OCR2 achieved higher performance while using the same number of tokens as Gemini 3 Pro
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
7/2/2026
First Seen
Valid until: 9/30/2026(expired)
Sources
DeepSeek OCR2, Kimi K2.5, and Microsoft Maia 200 All Launched on the Same Day
bilibiliinfinite灵感港
Related Claims
UnverifiedIn the OmniDocBench V1.5 benchmark, OCR2 achieved higher performance than Gemini 3 Pro while using the same number of tokens.83% similarUnverifiedDeepSeek V4 Pro was compared horizontally against flagship models such as Gemini 5.4.71% similarUnverifiedFrom a price-to-performance perspective, DeepSeek V4 completely crushes its competitors (OpenAI o4.6, Gemini 5.4, etc.).70% similarUnverifiedDeepSeek V4 Pro's Terminal Bench performance is on par with Claude Opus 4.7 and Gemini 3.1 Pro70% similarUnverifiedBaidu Qianfan released the end-to-end document intelligence model Qianfan OCR, with a parameter count of 4B, which scores slightly higher than DeepSeek OCR2 and also leads Gemini 3 Pro on multiple benchmark tests.66% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/55773API
curl https://kongchang.com/api/v1/knowledge/claims/55773MCP
get_claim(id=55773)