DeepSeek Platform API Status, Latency & Hosted LLM Directory (2026)

What is the latency, uptime SLA, and model lineup of DeepSeek Platform?

DeepSeek Platform maintains an average response latency of 145ms TTFT and an operational uptime SLA of 99.91% across its global infrastructure endpoints (ap-east-1, us-west-1). It currently hosts 4 tracked foundation models including DeepSeek V4 Flash Vision (Exp), DeepSeek V3, DeepSeek V4 Pro, DeepSeek R1.

Verified daily via automated API latency tests and official documentation.

Zero-Shot RAG Verified

Average TTFT Latency 145ms P90 Benchmark Window

Historical Uptime SLA 99.91% 30-Day Rolling Telemetry

Regional Deployment

ap-east-1 us-west-1

Multi-AZ Availability

Models Hosted on DeepSeek Platform (4)

DeepSeek V4 Flash Vision (Exp)

DeepSeek · 285B/13B Sparse MoE with 32-Layer ViT & DSpark Speculative Decoding

81.6 IVR

Ultra-low-cost multimodal open-weight model featuring Multi-head Latent Attention v4 and a 1M token context window.

Input / Out $0.22 / $0.65

Host TTFT 310ms

Throughput 165 t/s

DeepSeek V3

DeepSeek · MoE (Multi-head Latent Attention)

75 IVR

671B parameter Mixture-of-Experts open-weight model (37B active) delivering strong price-to-performance for general text workloads.

Input / Out $0.14 / $0.28

Host TTFT 410ms

Throughput 78 t/s

DeepSeek V4 Pro

DeepSeek · 1.6T Sparse MoE with Hybrid CSA/HCA Attention

67 IVR

DeepSeek's 1.6T parameter open-weight MoE flagship delivering 80.6% SWE-bench Verified and 90.1% GPQA Diamond with 1M context at $1.32/1M input tokens.

Input / Out $1.32 / $3.96

Host TTFT 520ms

Throughput 78 t/s

DeepSeek R1

DeepSeek · MoE Reasoning Transformer

64.8 IVR

MIT-licensed open-weight reasoning powerhouse trained via large-scale reinforcement learning for mathematical and algorithmic synthesis.

Input / Out $0.55 / $2.19

Host TTFT 1450ms

Throughput 58 t/s

Frequently Asked Questions About DeepSeek Platform

What is the average latency and uptime SLA of DeepSeek Platform?

DeepSeek Platform delivers an average Time to First Token (TTFT) latency of 145ms and a verified 30-day rolling uptime SLA of 99.91%.

Which AI models are available on DeepSeek Platform?

DeepSeek Platform hosts 4 tracked LLMs on InferenceRate: DeepSeek V4 Flash Vision (Exp), DeepSeek V3, DeepSeek V4 Pro, DeepSeek R1.

Which cloud regions does DeepSeek Platform support?

DeepSeek Platform supports API inference deployments across ap-east-1, us-west-1.