Parse: Benchmarks, Pricing & Context Window
Parse is a language model from Cohere, released in August 2026, with multimodal input, and a 8K-token context window.
Cohere Parse (`parse-v5.0`) is a 2.3B-parameter proprietary vision language model for high-volume enterprise document parsing. Proprietary north-micro-vision-instruct architecture (~4.6GB serving footprint); 8192 context. Converts PDF,
Parse benchmarks
Capability tiers
Standing within each category, adjusted for leaderboard depth.
Real tasks performance
High-confidence performance for Parse across real-world prompt categories. Only 95% intervals at most 4 points wide are shown.
Performance by conversation depth
How Parse holds up as conversations get longer.
Quality Tracker
Parse Performance Across Datasets
Scores sourced from the model's scorecard, paper, or official blog posts
llm-stats.com - Wed Sep 09 2026Parse context window
Input and output token limits for Parse, plus how it ranks on long-context understanding.
Try now
huggleMake it with
Parse.
Parse latency
Parse time to first token, sustained output throughput, and failed-request rate from live model usage over the trailing 7 days.
Parse examples
Recent arena outputs from Parse, picked from the highest-ranked matchups.
Parse license
Parse is a proprietary model available under its provider's product and API terms, has 2.3B parameters.
- License
- Proprietary
- Hosted access
- Parameters
- 2.3B
Proprietary license - usage restrictions apply
Parse resources
Official sources for Parse: provider documentation, official playground, official launch post.
FAQ
Common questions about Parse.