VECTOR WIREAI INTELLIGENCE
NVDA$1,847+3.2%MSFT$512+1.1%GOOGL$199-0.4%META$728+2.7%AMD$184-1.2%TSM$212+0.6%PLTR$98+4.1%AI IDX4,821+1.9%
PKT
SEEDRefresh Models Deals Regulatory Sources

ocrbench_v2

24 models tracked

OCRBench v2 — Explicit v2 of OCRBench (expanded task coverage); distinct version.

#ModelVendorBest scoreRunsLast seen
1Qwen3.7 Plus PreviewAlibaba67.112026-08-23
2NVIDIA Nemotron 3 Nano 30B A3BNVIDIA65.812026-05-22
3Qwen3.6 35B A3BAlibaba65.512026-07-02
4Qwen3.5 35B A3BAlibaba65.312026-07-02
5Gemini 3 ProGoogle63.412026-05-22
6NVIDIA Nemotron Nano 12B v2 VLNVIDIA61.212026-05-22
7Gemini 2.5 ProGoogle59.312026-05-22
8GLM-4.6V-Flash (9B)Z.ai5912026-05-22
9Qwen3.5 9BAlibaba58.712026-05-22
10Qwen2.5 Omni 7BAlibaba57.812026-08-23
11Llama 3.1 Nemotron Nano VL 8B V1NVIDIA56.412026-05-22
12GPT-5OpenAI55.532026-08-10
13Gemma 4 12BGoogle51.812026-07-02
14Gemini 1.5 ProGoogle51.612026-05-22
15GPT-5.2OpenAI50.512026-05-22
16Claude Opus 4.6Anthropic48.412026-05-22
17LFM2.5-VL-3BLiquid AI47.512026-08-14
18MiniCPM-o-4.5OpenBMB44.912026-05-22
19Gemma 4 E2BGoogle44.712026-08-12
20Molmo2-8BAllenAI44.712026-05-22
21LFM2-VL-3B (3.1B)Liquid AI43.912026-08-12
22Ministral 3 14BMistral43.312026-05-22
23MiniCPM-V 4.6OpenBMB40.412026-07-09
24Phi 4 Multimodal InstructMicrosoft38.122026-06-19

Best tracked score per model (default configuration; source-attributed and verification-tiered). Open a model for its full benchmark surface, provenance and pricing.