DeepSeek: DeepSeek V4 Pro 0813
Unavailable**In addition to DeepSeek, TencentCloud and AlibabaCloud now also offer the official 0731 version, while other providers are still using the old 0424 version**. . DeepSeek V4 Pro is DeepSeek’s flagship open-weight Mixture-of-Experts model, released in preview on April 24, 2026. It features 1.6 trillion total parameters, with 49 billion activated per token, and supports a one-million-token context window. Designed for advanced reasoning, coding, STEM tasks, and autonomous agents, it delivers state-of-the-art performance among open models and rivals leading proprietary systems. The model supports both thinking and non-thinking modes and is available through DeepSeek’s applications and APIs.
- By
- DeepSeek
- Input
- Text
- Output
- Text
- Context
- 1M tokens
- Max output
- 384K tokens
- Published
- 2026-08-12
- Capabilities
- ReasoningLong context
Providers
The provider that serves this model. Click the provider row to inspect live telemetry, pricing, and supported parameters.
| Provider | Input / 1M | Output / 1M | Context | Latency P50 | Throughput P50 | Success 24h |
|---|---|---|---|---|---|---|
 DeepSeek deepseek · offline | 69x Rp27.456/1M | 206x Rp82.368/1M | 1M | — | — | — |
Uptime
Share of requests that succeeded, per provider.
DeepSeekNo traffic| 06:00 PM | no traffic |
|---|---|
| 07:00 PM | no traffic |
| 08:00 PM | no traffic |
| 09:00 PM | no traffic |
| 10:00 PM | no traffic |
| 11:00 PM | no traffic |
| 12:00 AM | no traffic |
| 01:00 AM | no traffic |
| 02:00 AM | no traffic |
| 03:00 AM | no traffic |
| 04:00 AM | no traffic |
| 05:00 AM | no traffic |
| 06:00 AM | no traffic |
| 07:00 AM | no traffic |
| 08:00 AM | no traffic |
| 09:00 AM | no traffic |
| 10:00 AM | no traffic |
| 11:00 AM | no traffic |
| 12:00 PM | no traffic |
| 01:00 PM | no traffic |
| 02:00 PM | no traffic |
| 03:00 PM | no traffic |
| 04:00 PM | no traffic |
| 05:00 PM | no traffic |
Last 24 hours, per hour. Measured from requests through Morphic; none yet.
Throughput
P50 output tokens per second, whole request.
DeepSeek—| 06:00 PM | — |
|---|---|
| 07:00 PM | — |
| 08:00 PM | — |
| 09:00 PM | — |
| 10:00 PM | — |
| 11:00 PM | — |
| 12:00 AM | — |
| 01:00 AM | — |
| 02:00 AM | — |
| 03:00 AM | — |
| 04:00 AM | — |
| 05:00 AM | — |
| 06:00 AM | — |
| 07:00 AM | — |
| 08:00 AM | — |
| 09:00 AM | — |
| 10:00 AM | — |
| 11:00 AM | — |
| 12:00 PM | — |
| 01:00 PM | — |
| 02:00 PM | — |
| 03:00 PM | — |
| 04:00 PM | — |
| 05:00 PM | — |
Last 24 hours, per hour. Measured from requests through Morphic; none yet.
Latency
P50 time to the full response.
DeepSeek—| 06:00 PM | — |
|---|---|
| 07:00 PM | — |
| 08:00 PM | — |
| 09:00 PM | — |
| 10:00 PM | — |
| 11:00 PM | — |
| 12:00 AM | — |
| 01:00 AM | — |
| 02:00 AM | — |
| 03:00 AM | — |
| 04:00 AM | — |
| 05:00 AM | — |
| 06:00 AM | — |
| 07:00 AM | — |
| 08:00 AM | — |
| 09:00 AM | — |
| 10:00 AM | — |
| 11:00 AM | — |
| 12:00 PM | — |
| 01:00 PM | — |
| 02:00 PM | — |
| 03:00 PM | — |
| 04:00 PM | — |
| 05:00 PM | — |
Last 24 hours, per hour. Measured from requests through Morphic; none yet.
Activity
Tokens processed per day over the last 30 days.
| Oct 4 | 0 |
|---|---|
| Oct 5 | 0 |
| Oct 6 | 0 |
| Oct 7 | 0 |
| Oct 8 | 0 |
| Oct 9 | 0 |
| Oct 10 | 0 |
| Oct 11 | 0 |
| Oct 12 | 0 |
| Oct 13 | 0 |
| Oct 14 | 0 |
| Oct 15 | 0 |
| Oct 16 | 0 |
| Oct 17 | 0 |
| Oct 18 | 0 |
| Oct 19 | 0 |
| Oct 20 | 0 |
| Oct 21 | 0 |
| Oct 22 | 0 |
| Oct 23 | 0 |
| Oct 24 | 0 |
| Oct 25 | 0 |
| Oct 26 | 0 |
| Oct 27 | 0 |
| Oct 28 | 0 |
| Oct 29 | 0 |
| Oct 30 | 0 |
| Oct 31 | 0 |
| Nov 1 | 0 |
| Nov 2 | 0 |
Counts every request routed through Morphic.
Related models
More models from DeepSeek
DeepSeek: deepseek v4.1 LocalDeepSeek: DeepSeek V4 Flash 0731DeepSeek-V4-Flash is the efficiency-oriented variant of the DeepSeek V4 series, released as a preview and open-sourced alongside the flagship V4-Pro. It is designed for developers who need the V4 generation's long-context and reasoning capability at a faster, more economical API tier. Compared with V4-Pro, V4-Flash uses smaller total parameters and active parameters, resulting in faster response times and lower API cost. It retains reasoning capability close to V4-Pro and matches V4-Pro on simple agent tasks, with a measurable gap appearing only on the most demanding agent workflows. World knowledge is slightly below V4-Pro but remains competitive within the open-source landscape. Like V4-Pro, V4-Flash inherits the new attention mechanism built on token-dimension compression and DeepSeek Sparse Attention (DSA), supports a 1M context window as standard, and offers both thinking and non-thinking modes with a reasoning_effort parameter (high / max). Note on migration: the legacy model names deepseek-chat and deepseek-reasoner currently route to V4-Flash in non-thinking and thinking mode respectively, and will be retired on 2026-07-24. Existing integrations should migrate to the explicit model names deepseek-v4-flash or deepseek-v4-pro. The model is accessible via OpenAI ChatCompletions and Anthropic interfaces, with weights open-sourced on Hugging Face and ModelScope.
DeepSeek: DeepSeek V4Coding & reasoning model
DeepSeek: DeepSeek V4.1 Flash 1017Only DeepPeek and Alibaba Cloud providers are version 1017, while other providers are still old version. DeepSeek V4.1 Flash is a 552B-parameter MoE model built on a new causal encoder-decoder architecture, designed for higher capability, faster reasoning, higher throughput, and lower serving cost. It natively supports multimodal visual understanding and delivers flagship-level intelligence with significantly reduced KV cache requirements, making it well suited for high-throughput and cost-sensitive agentic workloads.
DeepSeek: DeepSeek V3.2DeepSeek-V3.2 is a reasoning-first large language model released by DeepSeek as the official successor to V3.2-Exp. It is designed with a focus on agentic capabilities and integrated reasoning for tool-use scenarios. If the request to the deepseek-reasoner model includes the tools parameter, the request will actually be processed using the deepseek-chat model. The model introduces a new large-scale agent training data synthesis method covering over 1,800 environments and 85,000+ complex instructions. DeepSeek-V3.2 is the first model from DeepSeek to integrate thinking directly into tool-use, supporting both thinking and non-thinking modes during tool interactions. According to DeepSeek's benchmarks, the model delivers performance comparable to GPT-5 level while balancing inference quality and output length. It is available via App, Web, and API, making it suitable for general-purpose daily use as well as complex agentic workflows.