Scenario guide
Best AI models for Multimodal Document Q&A
As of October 25, 2026, the highest-ranked model for Multimodal Document Q&A is GPT-5.4, with a scenario score of 91.0 out of 100. The score blends this scenario's benchmark weights with its cost weight.
Answering questions about scanned PDFs, charts, and screenshots. Relies on vision benchmarks plus reasoning.
Rankings use the same scenario weights and cost blending as the interactive leaderboard on AI Model Analyzer. Data is min-max normalised per benchmark; missing scores are skipped without penalty.
- 1GPT-5.4OpenAIScore 91.0Q 100.0In $1.25/M
- 2GPT-5.1OpenAIScore 90.5Q 97.8In $0.63/M
- 3Gemini 3.1 ProGoogleScore 90.2Q 98.4In $1.00/M
- 4GPT-5.2OpenAIScore 90.1Q 98.2In $0.88/M
- 5Gemma 4 31BGoogleScore 88.1Q 90.1In $0.20/M
- 6Gemini 2.5 FlashGoogleScore 83.2Q 84.4In $0.15/M
- 7Gemma 4 26B A4BGoogleScore 82.7Q 83.0In $0.15/M
- 8Gemini 2.5 ProGoogleScore 81.1Q 85.9In $0.63/M
- 9Claude Sonnet 4.6AnthropicScore 76.7Q 84.6In $3.00/M
- 10Claude Opus 4.5AnthropicScore 76.2Q 85.4In $5.00/M
- 11Claude Opus 4AnthropicScore 74.8Q 86.9In $15.00/M
- 12Claude Opus 4.6AnthropicScore 74.4Q 83.3In $5.00/M
- 13Gemini 2.0 FlashGoogleScore 71.5Q 67.9In $0.10/M
- 14Llama 4 MaverickMetaScore 71.3Q 69.4In $0.19/M
- 15Claude Sonnet 4AnthropicScore 70.0Q 76.2In $3.00/M