deepseek-v4-pro
ChatImageReasoningCodingLong-ContextMultimodal
DeepSeek V4 Pro is a large-scale MoE model with 1.6T total parameters, 49B active parameters, and a 1M-token context window. It is designed for advanced reasoning, coding, and long-running agent workflows, with strong performance in knowledge, math, and software engineering benchmarks. It suits full-codebase analysis, multi-step automation, and large-scale information synthesis.
Oken covers 26 routes, evaluated by model fingerprint, token usage (Caching), and uptime against the official baseline.
1routes
1routes
1routes
3routes
Route Quality
Provider Type
Billing Method
Route Performance
Uptime
Model Fingerprint
Platform Services
Payment Method
Supported Languages
Supported Regions
Recharge Platform