Inference

RealWorldQA

A grounded visual QA benchmark focused on answering practical questions about real-world images and scenes.

  • In index
  • Multimodal
ModelsModels with a result.12
Top resultBest result on this test.88.5%Qwen3.8-Flash-Next
Top-3 spreadPoints from first to third.0.8 pts
YearYear of release.N/A

Result and price

020406080100$0.3$1$3$10RESULTOUTPUT PRICE PER 1M TOKENS · LOG SCALE

Results

12 results
#ModelResultThe score from the source.
01Qwen3.8-Flash-Next88.5%
02Qwen3.8 Max88.0%
03Qwen3.8-Omni-Flash87.7%
04Qwen3.7 Plus86.9%
05Qwen3.8-27B85.9%
06Qwen3.6-35B-A3B85.3%
07Qwen3.6-27B84.1%
08Ternary Bonsai 2 27B80.1%
09LFM2.5-VL-3B73.1%
10ZAYA1-VL-8B65.0%
11North Micro Vision Instruct62.2%
12LFM2.5-VL-450M58.4%