Throughput Speed
75 tok/s
Model Overview & Capabilities
DeepSeek V4 Pro is a large-scale Mixture-of-Experts model from DeepSeek with 1.6T total parameters and 49B activated parameters, supporting a 1M-token context window. It is designed for advanced reasoning, coding,...
Technical Specifications
| Provider | deepseek |
| Context Window | 10,48,576 tokens |
| Output Price / 1M tokens | $1.89 |
| Academic Benchmark (MMLU) | 91.5% |
| License | Commercial API |
Explore LLM Pulse Interactive Tools
Frequently Asked Questions
Which AI model ranks highest for coding in 2026?
Coding rankings use reported third-party benchmark data. HumanEval values remain unverified unless metric-specific evaluator, model version, test methodology and test date evidence is available.
How is Pulse Score different from crowd-sourced leaderboards?
Pulse Score combines reported third-party benchmark data (HumanEval, MMLU, MATH) with real-world developer constraints: token throughput speed, input pricing per million tokens, and time-to-first-token latency.