Is DeepSeek V4 Pro shutting down? The short answer is just below — and it is good news if you are still running V4 Pro in production. I checked DeepSeek’s official API changelog today (23 September 2026) so you get the confirmed position, not the panic-thread version.
Short answer
- No — DeepSeek’s official changelog confirms V4 Pro API services continue after 14 September 2026, billing unchanged.
- The scare came from the 10 September release of V4.1 Flash and the rerouting of old Flash model names.
deepseek-flashnow calls V4.1 Flash;deepseek-v4-flashand the vision-exp name are temporarily routed to it.- V4.1 Flash brings native multimodal vision and reduced pricing; V4 Pro remains the agent-grade heavyweight.
- No new end-of-life date for V4 Pro has been published — but keep your stack model-agnostic anyway.
Is DeepSeek V4 Pro shutting down? What DeepSeek actually said
Straight from DeepSeek’s official API changelog (September 2026 entries, checked today): “In response to user demand, we have decided to continue providing API services for DeepSeek V4 Pro after September 14, 2026” — with the billing method remaining unchanged. That is a reversal of the planned wind-down, and it means production workloads on V4 Pro do not need an emergency migration.
So the direct answer to “is DeepSeek V4 Pro shutting down” is: no — the API stays up past the 14 September date, at the same billing, because enough users pushed back. DeepSeek has not published a new end-of-life date for it as of 23 September 2026.
Why everyone thought V4 Pro was being retired
The confusion was understandable, because three things happened almost at once (all per the official changelog):
- 10 September 2026: DeepSeek released DeepSeek-V4.1-Flash, the smallest model in its new architecture family, with native multimodal visual understanding and pricing reduced for the new model.
- Model name rerouting: the older names
deepseek-v4-flashanddeepseek-v4-flash-vision-expare temporarily routed to V4.1 Flash, anddeepseek-flashnow always calls the latest V4.1 Flash. - A 14 September cut-off date hanging over V4 Pro — the date after which service was in question until DeepSeek confirmed continuation.
Old model names being silently redirected is usually the tell that a family is being sunset — this time, the community reaction kept the heavyweight alive. If you hit odd behaviour during the transition, my note on DeepSeek’s server-busy errors covers the workarounds I use.
🔥 Want this set up without the guesswork? If you are running DeepSeek models for SEO content or agents and want your stack to survive exactly this kind of model-shuffle without downtime, that playbook exists. That is exactly what we do inside the AI Profit Boardroom — 3,700+ members, four live calls per week, daily tutorials, done-for-you templates and a 30-day roadmap.
Prefer to talk it through one-on-one first? Book a free SEO strategy session and we will map out your next steps together.
DeepSeek V4 Pro vs V4.1 Flash: which should you run now?
| DeepSeek V4 Pro | DeepSeek V4.1 Flash | |
|---|---|---|
| Status | API continues past 14 Sept 2026, billing unchanged | Released 10 Sept 2026, current flagship of the new family |
| Positioning | Heavyweight; August GA emphasised agent capabilities and native Responses API support | Smallest model of the new architecture; built for speed and throughput |
| Vision | — | Native multimodal visual understanding |
| API name | V4 Pro endpoints unchanged | deepseek-flash (old Flash names temporarily routed here) |
| Billing | Unchanged; peak/off-peak with off-peak at 50% of peak | Pricing reduced with the new release |
My take: keep agentic, complex workloads on V4 Pro for now, and move high-volume, latency-sensitive or vision jobs to V4.1 Flash. I have broken down the new model’s vision capabilities and its API pricing separately if you are costing a migration.
What this means for your API costs
Nothing changes on the V4 Pro side — that is the whole point of the announcement. The billing method stays as it was, including the peak/off-peak pricing DeepSeek introduced with V4 Pro’s general availability in August, where off-peak calls run at 50% of peak rates. If your pipelines can batch work into off-peak windows, that discount is still the single easiest cost lever on the platform. Meanwhile V4.1 Flash arrived with reduced pricing, so the cheap tier got cheaper — a squeeze play on every rival’s small model.
Model churn like this is exactly why I keep every client build model-agnostic. It is one of the core systems inside the AI Profit Boardroom, and if you want help pressure-testing your own setup you can book a free SEO strategy session and we will walk through it together.
Is DeepSeek V4 Pro shutting down? The bottom line
No. As of 23 September 2026, DeepSeek’s own changelog confirms V4 Pro API services continue past the 14 September cut-off with billing unchanged — a genuine user-demand reversal. Treat it as a reprieve, not a guarantee: the new-architecture family that V4.1 Flash opened is clearly where DeepSeek is heading, so enjoy V4 Pro while it lasts and keep your integration one config line away from a swap. I will update this page if DeepSeek publishes any new dates.
FAQ: is deepseek v4 pro shutting down
Is the DeepSeek V4 Pro API being discontinued?
No. DeepSeek’s official changelog states that, in response to user demand, API services for DeepSeek V4 Pro continue after 14 September 2026, with the billing method unchanged.
Why did people think DeepSeek V4 Pro was shutting down?
The release of DeepSeek-V4.1-Flash on 10 September 2026, plus the rerouting of older Flash model names to the new model, read like a sunset plan. DeepSeek then confirmed V4 Pro stays.
What happened to the old DeepSeek Flash model names?
Per the changelog, deepseek-v4-flash and deepseek-v4-flash-vision-exp are temporarily routed to V4.1 Flash, and deepseek-flash now calls the latest V4.1 Flash model.
Does DeepSeek V4.1 Flash support vision?
Yes — DeepSeek says V4.1 Flash natively supports multimodal visual understanding, replacing the separate experimental vision model from August.
Is DeepSeek V4 Pro pricing changing?
DeepSeek says the billing method remains unchanged. V4 Pro kept the peak/off-peak structure introduced at its August GA, with off-peak at 50% of peak rates.
Should I migrate from V4 Pro to V4.1 Flash?
Not automatically. V4 Pro remains the agent-focused heavyweight; V4.1 Flash is the fast, cheap multimodal option. Test your actual workloads on both before moving anything.
Related reading
- DeepSeek V4.1 Flash vision capabilities explained
- DeepSeek V4.1 Flash API pricing: full cost breakdown
- DeepSeek server busy errors: fixes that work
Next steps: V4 Pro lives on — but the smart move is a stack that does not care which model wins. Join 3,700+ members inside the AI Profit Boardroom for the live calls, daily tutorials and done-for-you templates, or book a free SEO strategy session and get a personal plan for your site.
About the author
Julian Goldie is an SEO agency owner with 10+ years in SEO, 394K+ subscribers on YouTube, a 100% job-success score on Upwork, 75K+ community members across his groups, and the author of a best-selling SEO book. He runs the AI Profit Boardroom community, and you can book a free SEO strategy session with his team any time.
Last updated September 2026. This is the living guide to is deepseek v4 pro shutting down — it gets updated as the tools change.
