DeepSeek V4 Pro
DeepSeek · open weights · served by 83 providers · tool calling, reasoning
deepseek/deepseek-v4-pro
- Weights
- Open
- Released
- 2026-04-24
- Knowledge cutoff
- 2025-05
- Context window
- 1M
When each platform deprecates DeepSeek V4 Pro, as LiteLLM records it. Plan the move before the date: after it, the platform may stop serving the model or stop updating it.
- wandbdeprecated 2026-10-05wandb/deepseek-ai/DeepSeek-V4-Prosuccessor: wandb/deepseek-ai/DeepSeek-V4-Pro-0813
- Basetendeprecated 2026-10-16baseten/deepseek-ai/DeepSeek-V4-Prosuccessor: baseten/deepseek-ai/DeepSeek-V4-Pro-0813
deepseek-ai/DeepSeek-V4-Pro at commit b5968e9190, last changed 2026-06-22
- Licence
- mit
- Access
- Open download
- Library
- transformers
- Formats
- .safetensors
- No model signature file is published with the weights.
What the Hub's scanners report
22 weight files, 22 with a scan result. No scanner reports an issue with them.
Facts from the Hugging Face Hub and its file scanners (Protect AI, Cisco, Hugging Face picklescan, VirusTotal, JFrog), as of 2026-11-18. ShieldMap does not copy model cards.
- Abacusdeepseek-ai/DeepSeek-V4-Prodeprecated
- Adzuki AIdeepseek-v4-pro
- ai&deepseek-ai/deepseek-v4-pro
- AIHubMixalicloud-deepseek-v4-prodeprecated
- AIHubMixdeep-deepseek-v4-prodeprecated
- AI/ML APIdeepseek/deepseek-v4-pro
- Alibaba (China)deepseek-v4-pro
- Alibaba Token Plandeepseek-v4-pro
- Alibaba Token Plan (China)deepseek-v4-pro
- AnyAPIdeepseek/deepseek-v4-pro
- APIKOdeepseek-v4-pro
- Arceedeepseek/deepseek-v4-prodeprecated
- Aurikodeepseek-v4-pro
- Azuredeepseek-v4-pro
- Azure Cognitive Servicesdeepseek-v4-pro
- BytePlus ModelArkdeepseek-v4-pro-260425
- ClinePasscline-pass/deepseek-v4-pro
- Cloudflare AI Gatewaydeepseek/deepseek-v4-pro
- CommandCodedeepseek/deepseek-v4-pro
- Cortecsdeepseek-v4-pro
- CrossModeldeepseek/deepseek-v4-pro
- Deep Infradeepseek-ai/DeepSeek-V4-Pro
- DigitalOceandeepseek-v4-pro
- DotHooshdeepseek-v4-pro
59 more providers
EBCloud (DeepSeek-V4-Pro), Eden AI (deepseek/deepseek-v4-pro), EmpirioLabs AI (deepseek-v4-pro), Ezkielyna (deepseek-v4-pro), FastRouter (deepseek/deepseek-v4-pro), FrogBot (deepseek-v4-pro), Gisoo (deepseek-v4-pro), GMI Cloud (deepseek-ai/DeepSeek-V4-Pro), HPC-AI (deepseek/deepseek-v4-pro), Hugging Face (deepseek-ai/DeepSeek-V4-Pro), Impossibl (deepseek/deepseek-v4-pro), Inference APIs (deepseek-ai/DeepSeek-V4-Pro), Jalapeno Cloud (DeepSeek-V4-Pro), JZSTOKEN (deepseek-v4-pro), Kenari (deepseek-v4-pro), Kilo Gateway (deepseek/deepseek-v4-pro), DevPass (LLM Gateway) (deepseek-v4-pro), LLM Gateway (alibaba/deepseek-v4-pro), LLM Gateway (baidu/deepseek-v4-pro), LLM Gateway (bytedance/deepseek-v4-pro), LLM Gateway (deepinfra/deepseek-v4-pro), LLM Gateway (deepseek/deepseek-v4-pro), LLM Gateway (runware/deepseek-v4-pro), LLM Gateway (tencent/deepseek-v4-pro), LLM Gateway (together-ai/deepseek-v4-pro), Merge Gateway (deepseek/deepseek-v4-pro), Moark (DeepSeek-V4-Pro), Model Oracle AI (deepseek-v4-pro), Modelis (deepseek-v4-pro), Novita AI (deepseek/deepseek-v4-pro), Ofox (deepseek/deepseek-v4-pro), Ollama Cloud (deepseek-v4-pro), OpenCode (deepseek-v4-pro), OpenCode Go (deepseek-v4-pro), OpenRouter (deepseek/deepseek-v4-pro), Opper (deepseek-v4-pro), OrcaRouter (deepseek/deepseek-v4-pro), Poe (empiriolabs/deepseek-v4-pro-el), Prime Inference (deepseek/deepseek-v4-pro), Requesty (deepseek-v4-pro), Requesty (deepseek-v4-pro@eu), routing.run (deepseek-v4-pro), SandBase (deepseek/deepseek-v4-pro), SCNet Token Plan (DeepSeek-V4-Pro), SCX.ai (DeepSeek-V4-pro), SiliconFlow (deepseek-ai/DeepSeek-V4-Pro), SiliconFlow (China) (deepseek-ai/DeepSeek-V4-Pro), Standard Compute (deepseek/deepseek-v4-pro), TensorX (deepseek/deepseek-v4-pro), TokenGo (deepseek/deepseek-v4-pro), UnoRouter (deepseek-v4-pro), UnoRouter (deepseek-v4-pro:free), Venice AI (deepseek-v4-pro), Vercel AI Gateway (deepseek/deepseek-v4-pro), Vivgrid (deepseek-v4-pro), Volcengine Ark Coding Plan (deepseek-v4-pro), Wally (deepseek-v4-pro-fast), Watt Cloud (DeepSeek-V4-Pro), ZenMux (deepseek/deepseek-v4-pro)
Open weights are loaded and served by code that has its own advisories. These inference servers and model loaders are on ShieldMap; load weights only from the publisher's own repository, and prefer safetensors over formats that can run code.
- vllm · 131 advisories
- transformers · 24 advisories
- keras · 20 advisories
- onnx · 17 advisories
- bentoml · 16 advisories
- github.com/ollama/ollama · 16 advisories
- torch · 13 advisories
- lmdeploy · 10 advisories
- sglang · 10 advisories
- skops · 4 advisories
Sources: models.dev (MIT) and the LiteLLM model catalogue (MIT). Deprecation dates are as each platform published them and as LiteLLM records them; check the platform before you rely on one. Data as of 2026-11-22.