DeepSeek-V4-Pro at a glance
DeepSeek V4 arrived as an open-weight preview with a 1.6T-parameter Pro model and a 284B Flash model; V4-Pro reached general availability on August 13, 2026.
What changed with DeepSeek-V4-Pro
On April 24, 2026, DeepSeek released a preview of V4: DeepSeek-V4-Pro (1.6T total, 49B active parameters) and DeepSeek-V4-Flash (284B total, 13B active). Both are open-sourced on Hugging Face, support thinking and non-thinking modes, and make 1M tokens the default context across DeepSeek services.
The API accepts both OpenAI ChatCompletions and Anthropic formats, and DeepSeek said its old deepseek-chat and deepseek-reasoner names would be retired after July 24, 2026. On August 13, 2026, V4-Pro reached general availability with agent upgrades, low/high/max reasoning effort, Responses API support, and off-peak pricing at half the peak rate.
DeepSeek-V4-Pro benchmarks and pricing
Values are copied from the linked source on the date shown. They are not IA Compa measurements.
- Artificial Analysis Intelligence Indexas of · Source for Artificial Analysis Intelligence Index: artificialanalysis.ai (opens in a new tab)
- 23 (v5.0.1)
- Output speedas of · Source for Output speed: artificialanalysis.ai (opens in a new tab)
- 72 tokens/s
- Input price per 1M tokensas of · Source for Input price per 1M tokens: artificialanalysis.ai (opens in a new tab)
- $1.32
- Output price per 1M tokensas of · Source for Output price per 1M tokens: artificialanalysis.ai (opens in a new tab)
- $3.96
- Context windowas of · Source for Context window: api-docs.deepseek.com (opens in a new tab)
- 1M tokens
How it compares
Other models in the same category that we have covered, using each model's latest recorded value. (Models)
Where to compare DeepSeek-V4-Pro
Independent leaderboards and official resources for comparing this model yourself. Rankings change often; each site shows its own current data.
Leaderboards
General leaderboards
Why DeepSeek-V4-Pro matters for businesses
Artificial Analysis lists the GA version (V4 Pro 0813, Max) at 23 on Intelligence Index v5.0.1, about 72 tokens per second, and $1.32/$3.96 per 1M input/output tokens. Teams using the legacy model names needed to migrate; the open weights also make V4 a self-hosting candidate.
Sources for this story
- DeepSeek V4 Preview Release (opens in a new tab)Official
DeepSeek API Docs · api-docs.deepseek.com
- DeepSeek-V4-Pro GA Release (opens in a new tab)Official
DeepSeek API Docs · api-docs.deepseek.com
- Change Log (opens in a new tab)Official
DeepSeek API Docs · api-docs.deepseek.com
- DeepSeek V4 Pro model analysis (opens in a new tab)Benchmark
Artificial Analysis · artificialanalysis.ai
Last verified: . IA Compa summarizes public announcements for business readers. We are not affiliated with the companies mentioned. Always confirm details, pricing, and availability with the official source.
- #llm
- #deepseek
- #open-weights
- #long-context