deepseek-v4-flash
ChatImageReasoningCodingLong-ContextMultimodal
DeepSeek V4 Flash is an efficiency-optimized MoE model with 284B total parameters, 13B active parameters, and a 1M-token context window. It is built for fast inference and high-throughput workloads while retaining strong reasoning and coding. Its hybrid attention handles long context efficiently, making it ideal for coding assistants, chat systems, and cost-sensitive agent workflows.
Oken covers 26 routes, evaluated by model fingerprint, token usage (Caching), and uptime against the official baseline.
2routes
0routes
0routes
4routes
Route Quality
Provider Type
Billing Method
Route Performance
Uptime
Model Fingerprint
Platform Services
Payment Method
Supported Languages
Supported Regions
Recharge Platform