DeepSeekAIWS category: Reasoning & CodingOpen Weights (MIT)
DeepSeek V4 Pro
DeepSeek V4 Pro, an API model supporting thinking and non-thinking modes.
Source-backed facts
Field verified
Verification
Field verified
Profile
Source-backed
Context
1M tokens
Max output
384K tokens
API
Available
Specifications
Context
1M tokens
Max output
384K tokens
CompanyDeepSeek
Release DateApr 24, 2026
Provider statusNot published
LicenseMIT
API availabilityAvailable
Capability Map
Verified fieldsReasoning
Supported
API access
Available
Open weights
Open weights
Native input modalities
Not published
Native output is what the model returns directly. Tool capabilities are separate.
Developer API Rates
Verified public rateInput / 1M
$1.32
Output / 1M
$3.96
Cached $0.0440 / 1M
Displayed rates use peak-hour pricing: input cache miss $1.32, cache hit $0.044, output $3.96 per 1M tokens.
Off-peak rates are $0.66 input cache miss, $0.022 cache hit and $1.98 output per 1M tokens.
Peak hours: Monday–Friday 01:00–04:00 and 06:00–10:00 UTC. All other hours are off-peak. The calculator uses peak rates and does not schedule traffic.
Verified Capabilities
- 1M tokens published context window
- 384K tokens published maximum output
- Reasoning support listed by the provider
- Direct API availability is published
Pricing & Availability Notes
- Displayed rates use peak-hour pricing: input cache miss $1.32, cache hit $0.044, output $3.96 per 1M tokens.
- Off-peak rates are $0.66 input cache miss, $0.022 cache hit and $1.98 output per 1M tokens.
- Peak hours: Monday–Friday 01:00–04:00 and 06:00–10:00 UTC. All other hours are off-peak. The calculator uses peak rates and does not schedule traffic.
Provider API model ID
Use this identifier in the provider API where supported.
deepseek-v4-pro