AI Research
DeepSeek V4 Pro Reaches GA with Adjustable Reasoning and Responses API Support
DeepSeek has released the GA version of V4 Pro for its app, web service and API. The 0813 model adds low, high and max reasoning effort, native Responses API support, and a new peak and off-peak pricing schedule from 16 August.
AI-narrated summary, not a word-for-word reading.
Key takeaways
- check_circleDeepSeek announced V4 Pro GA on 13 August 2026 and made it available in Expert Mode on app and web, as well as through its API.
- check_circleThe GA version supports low, high and max reasoning effort, plus native OpenAI Responses API support aimed at Codex workflows.
- check_circleAIMI first saw Ollama Cloud list deepseek-v4-pro:preview at 00:01:27 SAST on 14 August. That is a provider route observation, not the GA release time.
From the April preview to the 0813 GA build
DeepSeek first released the V4 preview on 24 April 2026. On 13 August, the company announced the general availability build now identified in its documentation as DeepSeek-V4-Pro-0813.
The API model name remains deepseek-v4-pro, so existing callers do not need to switch to a dated model ID. DeepSeek says the update improves agent performance in production, but that is a vendor claim until independent evaluations test the GA build on comparable workloads.
Reasoning control and agent integrations
The GA release adds adjustable reasoning effort to V4 Pro and V4 Flash. DeepSeek maps low effort to simpler tasks, high to everyday agent work, and max to more complex jobs.
DeepSeek also added native support for the OpenAI Responses API and published a Codex setup path. This gives agent tools another integration option beyond the OpenAI-compatible chat completions and Anthropic interfaces already documented for V4.
Where the GA model is available
DeepSeek lists V4 Pro in Expert Mode on its app and website, and through the official API as deepseek-v4-pro. The company documentation now points that stable API alias to the 0813 build.
AIMI also tracks V4 Pro routes through OpenRouter, OpenCode Go, Cline and Ollama Cloud. Provider labels do not always match DeepSeek's release naming. In particular, Ollama Cloud surfaced deepseek-v4-pro:preview at 22:01:27 UTC on 13 August, or 00:01:27 SAST on 14 August. That route confirms provider availability only. It does not change the maker's 13 August GA date or prove that the provider route serves the 0813 build.
A pricing change follows on 16 August
DeepSeek will introduce separate peak and off-peak API rates at 16:00 UTC on 16 August 2026. The company says off-peak rates will be 50% below peak pricing.
That pricing change is scheduled, not yet active at the time of publication. Teams moving recurring work to V4 Pro should check DeepSeek's live pricing page after the change takes effect rather than hard-coding announcement figures into cost models.
The release timeline needs careful labels
The April date belongs to the V4 preview. The 13 August date belongs to DeepSeek's GA announcement and 0813 model build. AIMI's 14 August SAST observation belongs to an Ollama Cloud endpoint listing.
DeepSeek's announcement gives a calendar date but no precise publication time. That means an exact elapsed-time comparison between the maker announcement and provider sightings would be false precision. There were no earlier verified releases in this AIMI package's 14 August SAST batch, so this is not being framed as part of a same-day release race.
Frequently asked questions
Is DeepSeek V4 Pro still a preview?
No. DeepSeek announced V4 Pro GA on 13 August 2026. Some provider routes still contain the word preview, which describes that provider label rather than the current maker release status.
What API model name should developers use?
DeepSeek says the API name remains deepseek-v4-pro. Its documentation identifies the current model behind that alias as DeepSeek-V4-Pro-0813.
Which reasoning effort levels are supported?
DeepSeek documents low, high and max effort for V4 Pro and V4 Flash.
When does the new API pricing start?
The peak and off-peak schedule takes effect at 16:00 UTC on 16 August 2026. DeepSeek says off-peak rates will be 50% lower than peak rates.
Sources
Copyright & image credits
The article image is credited to © DeepSeek, used for news reporting. Source: DeepSeek: DeepSeek-V4-Pro GA Release.