All models
DeepSeek
DeepSeek V4 Pro 0423 frontierruns locally
deepseek/deepseek-v4-pro
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,...
Input / 1M
$1.32
USD per million tokens
Output / 1M
$3.96
USD per million tokens
Est. cost / token
$0.00000198
$1.98 per 1M, 3:1 mix
Context window
1M
max output 393K
Major features
Open weightsCachingReasoningJSON schemaTools
Task suitability (heuristic 1–5, not a benchmark)
- coding5
- reasoning5
- math5
- summarization5
- extraction5
- classification5
- chat5
- creative5
- translation5
- vision1
- long context5
- agentic5
Run it locally
Will it run on my machine? →| Quantization | Quality | Memory needed (8K ctx) |
|---|---|---|
| Q2_K | poor | 714 GB |
| Q3_K_M | fair | 831 GB |
| Q4_K_Mrecommended | good | 1037 GB |
| Q5_K_M | very good | 1205 GB |
| Q6_K | excellent | 1387 GB |
| Q8_0 | near-lossless | 1794 GB |
| F16 | lossless | 3368 GB |
Estimates.
Price history
| Recorded | Input | Output | Cached | Source |
|---|---|---|---|---|
| 2026-11-05 03:20 | $1.32 | $3.96 | $0.044 | litellm+openrouter |
A row is added whenever a sync detects a price change. Vendor pricing source
Pricing (USD per 1M tokens)
- Input
- $1.32
- Output
- $3.96
- Cached input
- $0.044
- Blended (3:1)
- $1.98
Details
- Company
- DeepSeek
- Version
- 4
- Family
- DeepSeek
- Released
- 2026-04-24
- Parameters
- 1.6T
- Max output
- 393K
- Open weights
- Yes
- Inputs
- text
- Last synced
- 21 min ago