DeepSeek V4 Pro 0813 is a large-scale mixture-of-experts model from DeepSeek. This is the GA release of DeepSeek V4 Pro.
Tested on 21 benchmarks · BenchGecko score 66.8. Top scores: OTIS Mock AIME 2024-2025 (98.6%), ARC-AGI (90.5%), Dtbench (89.8%).
Unusual and adversarial machine learning challenges. Tests robustness of reasoning about edge cases in ML systems.
Abstraction and Reasoning Corpus. Tests fluid intelligence through novel visual pattern recognition puzzles. Core measure of general intelligence.
ARC-AGI 2, harder sequel to ARC. More complex abstract reasoning patterns that test generalization ability beyond training data.
Mock AIME (American Invitational Mathematics Exam) problems from OTIS. Tests mathematical competition performance.
- Typetext
- Context1.0M tokens (~524 books)
- ReleasedAug 2026
- LicenseProprietary
- StatusActive
- Cost / Message~$0.006
Frequently Asked Questions
Key facts · as of 2026-10-05
- DeepSeek V4 Pro 0813 by DeepSeek. BenchGecko score 66.8, rank 49 of 312 scored models (normalized average of public benchmark scores).
- List price $0.55 input · $5.00 output per 1M tokens (as of 2026-10-05).
- Sold by 20 providers (as of 2026-10-05): Relace (fp4) $0.19 in / $4.20 out · Ionstream $0.24 in / $1.98 out · Wafer $0.45 in / $5.00 out · StreamLake $0.66 in / $1.98 out · Novita (fp8) $0.99 in / $2.97 out · and 15 more. Every provider
- Gecko Tests: Who Are You E (Sometimes says it is OpenAI or Anthropic) · Tokenizer Tax A (17% more tokens outside English).
How to cite · data as of 2026-10-05
DeepSeek V4 Pro 0813 · benchmarks, pricing and providers. BenchGecko, data as of 2026-10-05. https://benchgecko.ai/model/deepseek-v4-pro-0813
Credit "Source: BenchGecko" with a link. Prices per provider and Gecko Tests are BenchGecko data (CC BY 4.0); benchmark scores keep their original source, listed in the JSON. JSON · llms.txt · MCP