DeepSeek
DeepSeek V4 Pro
As of 2026-11-15, DeepSeek V4 Pro costs $1.32 per 1 million input tokens and $3.96 per 1 million output tokens at the official list price (DeepSeek API). The lowest input price recorded on any channel is $0.90 (OpenRouter). Source: AuraMatrix Token Price Index; snapshot SHA-256 5126b5f69e78f6da….
Official list price · DeepSeek API
- Input
- 1.32
- Output
- 3.96
- Cached input
- 0.044
- Cache write
- —
USD per 1M tokens. A dash means the price was not recorded, not that it is free.
Off-peak hours (see source): input 0.66, output 1.98, cached input 0.022.
Billing rule: "Off-peak hours (see source)" applies outside 01:00–04:00 and 06:00–10:00 UTC, Monday to Friday. Chinese public holidays are off-peak in full. VERIFIED
Input price is the cache-miss rate.
Observed 2026-11-13 22:15 UTC. The source still stated these prices at the last daily check, 2026-11-15 22:15 UTC.
Cheaper in the same tier
How to read this
Official list prices, blended 3 parts input to 1 part output, against models the vendors place in the same tier. Price only: a cheaper model is not an equal one, and nothing here measures quality.
What this price buys
Estimates from the official list price above: tokens read and written for a typical job of that size, times the price. Your own usage will differ. See how these are worked out, and try other jobs →
All channels
| Channel | Type | Region | Input | Output | Cached in | vs officialinput / output | Status | Source |
|---|---|---|---|---|---|---|---|---|
| DeepSeek APIDeepSeek · observed 2026-11-13 22:15 UTC · confirmed 2026-11-15 22:15 UTCPeak rate shown as the list price; off-peak rate is a separate tier.Input price is the cache-miss rate. | Officialpeak / off-peak pricing by the operator | globalGlobal endpoint | 1.32 | 3.96 | 0.044 | reference | VERIFIED | Source |
| OpenRouterlowest inputOpenRouter · observed 2026-11-13 22:15 UTC · confirmed 2026-11-15 22:15 UTCPrice listed by OpenRouter's public models API. The billed price can depend on the provider OpenRouter routes to.OpenRouter model id deepseek/deepseek-v4-pro. Converted from per-token prices. | Aggregatormarketplace's own price | globalGlobal endpoint | 0.90 | 1.50 | 0.25 | −32% / −62% | LISTED | Source |
China sites, priced in CNY
| Channel | Plan | Input | Output | Cached in | vs standardinput / output | Status | Source |
|---|---|---|---|---|---|---|---|
| DeepSeek API, CNY price listDeepSeek · observed 2026-11-13 22:15 UTC · confirmed 2026-11-15 22:15 UTCOff-peak hours (see source): input ¥4.50, output ¥13.50Input price is the cache-miss rate. | Standard | ¥9.00 | ¥27.00 | ¥0.30 | standard | VERIFIED | Source |
Measured
One small request an hour, plus one longer request a day to measure speed, sent from our Los Angeles server over the last 7 days. This does not test model quality; it shows whether the API answered, how soon and how fast. All models →
| Channel | Answered | First tokenmedian | Speedtokens per second | Charged vs list price | Last call |
|---|---|---|---|---|---|
| DeepSeek API | 100%64 of 64 calls | 1.4 s | 474 | — | 2026-11-15 22:07 UTC |
| OpenRouterserved by Relace, Baidu, StreamLake | 100%107 of 107 calls | 1.8 s | 499 | −13%111 requests | 2026-11-15 22:07 UTC |
Price history
Input price per 1M tokens on each channel, from the first record to the last daily check. A price holds until the next record that states another one, so the lines are steps; a flat line means the price has not moved since we started recording.
| Channel | Since | First | Course | Now | Moves |
|---|---|---|---|---|---|
| OpenRouter | 2026-11-03 | 1.60 | 0.90−44% | 11last 2026-11-13 22:15 UTC | |
| DeepSeek API | 2026-11-03 | 1.32 | 1.32 | · | |
| DeepSeek API, CNY price list | 2026-11-09 | ¥9.00 | ¥9.00 | · |
Names this model is listed under
- deepseek-v4-pro
- DeepSeek-V4-Pro-0813
- deepseek/deepseek-v4-pro
Compare with
- DeepSeek V4 Pro vs Qwen3.8 Max
- DeepSeek V4 Pro vs Claude Fable 5.1
- DeepSeek V4 Pro vs Claude Fable 5.2
- DeepSeek V4 Pro vs Claude Mythos 5.2
- DeepSeek V4 Pro vs Claude Opus 5.5
- DeepSeek V4 Pro vs Gemini 4 Argon
- DeepSeek V4 Pro vs MiniMax M3
- DeepSeek V4 Pro vs Kimi K3
- DeepSeek V4 Pro vs Kimi K3.1 Pro
- DeepSeek V4 Pro vs GPT-6 Astra
- DeepSeek V4 Pro vs GLM-5.3
Check a pricePaying for this model, or holding a quote?See where your price stands →
In your AIAsk your assistant: “Where is DeepSeek V4 Pro cheapest right now, and how fast is it?”Connect in one click →
Sources
Each price above was read from one of these snapshots. The hash identifies the exact page content that was seen.
OpenRouter public models API
- URL
- https://openrouter.ai/api/v1/models
- Fetched
- 2026-11-13 22:15 UTC
- SHA-256
- 22716c39df25cb73975f808861d4f7559e4c15bd6f14bb8583ada56bdf46cfd3
- Size
- 850,292 bytes
- Records
- q-20261113-openrouter-deepseek-v4-pro · OpenRouter · read from a machine-readable price list by reobserve.py (script) · reviewed by codex (agent) · OpenRouter model id deepseek/deepseek-v4-pro. Converted from per-token prices.
DeepSeek API: Models & Pricing
- URL
- https://api-docs.deepseek.com/quick_start/pricing/
- Fetched
- 2026-11-13 22:15 UTC
- SHA-256
- 5126b5f69e78f6dacb0a8bfc91e0cf0b705416b0b918b12e1f45f66b122e7741
- Size
- 29,133 bytes
- Records
- q-20261113-deepseek-api-deepseek-v4-pro · DeepSeek API · read from a machine-readable price list by claude-opus-5-5 (headless reader) · reviewed by codex (agent) · Input price is the cache-miss rate.
DeepSeek API: 模型 & 价格 (Chinese page, CNY)
- URL
- https://api-docs.deepseek.com/zh-cn/quick_start/pricing/
- Fetched
- 2026-11-13 22:15 UTC
- SHA-256
- 23b70cda2e51dde6685a59c7336d0cf7a672b54f9000004b9166514f5e81510e
- Size
- 29,314 bytes
- Records
- q-20261113-deepseek-api-cny-deepseek-v4-pro · DeepSeek API, CNY price list · read from a machine-readable price list by claude-opus-5-5 (headless reader) · reviewed by codex (agent) · Input price is the cache-miss rate.