The Orchestrator › Models › DeepSeek

DeepSeek: DeepSeek V4 Pro 0423

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,...

Specifications

ProviderDeepSeek
Model IDdeepseek/deepseek-v4-pro
Context window1.0M tokens (1,048,576)
Modalitiestext → text
TokenizerDeepSeek
Router sourcesOpenRouter, TrustedRouter, Vercel AI Gateway, Cloudflare
Released2026-04-24
Tracked since2026-11-14
Popularity#42 by usage · peak 2.7% spend share in an OpenRouter category
ReasoningSupported

Pricing (per million tokens)

VariantSourceAPI IDInputOutputContext
standardOpenRouterdeepseek/deepseek-v4-pro$0.9$1.501.0M
standardTrustedRouterdeepseek/deepseek-v4-pro$0.158$0.6331.0M
standardVercel AI Gatewaydeepseek/deepseek-v4-pro$1.32$3.961.0M
standardCloudflaredeepseek/deepseek-v4-pro$1.74$3.48—

Recorded price history

DateInput / 1MOutput / 1MContext
2026-11-14$1.74$3.48—
2026-11-14$1.32$3.961.0M
2026-11-14$0.158$0.6331.0M
2026-11-14$0.9$1.501.0M

Capabilities

Agentic tool useStructured outputs1M+ context windowControllable reasoningNo content storage (attested)Zero data retention

Trust flags come from TrustedRouter's per-provider attestations.

Supported parameters

frequency_penalty, include_reasoning, logit_bias, logprobs, max_completion_tokens, max_tokens, min_p, presence_penalty, reasoning, reasoning_effort, repetition_penalty, response_format, seed, stop, structured_outputs, temperature, tool_choice, tools, top_k, top_logprobs, top_p