Cohere Parse 5 loses the benchmark on points. It wins on cost per page.

Cohere’s Parse 5, a 2.3-billion-parameter vision-language model, converts PDFs, slides and images into structured Markdown at $1.50 per 1,000 pages, supporting nine major languages and offering both plain-Markdown and block-mode outputs with HTML tables and bounding-box data. In Cohere’s ParseBench, Parse 5 scores 79.2—below GPT-5.5 (84.4), Opus 4.8 (84.3) and Gemini 3.5 Flash (81.8) but above LlamaParse’s Cost-Effective tier (78.3); it omits layout and chart extraction, focusing on reading-order Markdown and chart descriptions.