DeepSeek · Released Aug 12, 2026DeepSeek V4 Pro 0813
DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes.
- Capability
- 155.4ECI · #26 of 148
- Input
- $0.66per 1M tokens
- Output
- $1.98per 1M tokens
- Context
- 1M384K max output
What it costs
- Input
- $0.66per million tokens
- Output
- $1.98per million tokens
- Cached input
- $0.022per million tokens
- Blended (3:1)
- $0.99pricier than 54% of priced models
Typical monthly bills
| Monthly usage | Estimated cost |
|---|---|
| Side project2M input + 0.5M output tokens | $2.31 |
| Team assistant25M input + 5M output tokens | $26.40 |
| Production app250M input + 50M output tokens | $264.00 |
Official DeepSeek API price, as listed on models.dev.
Independent benchmarks
#26 of 148 scored models
The shaded band is Epoch AI’s confidence range (153.7–157.6); the tick marks the median scored model.
Scores from Epoch AI, run independently of DeepSeek.
The details
- Lab
- DeepSeek
- Released
- Aug 12, 2026
- Context window
- 1,000,000 tokens
- Max output
- 384,000 tokens
- Inputs
- Text
- Output
- Text
- Reasoning
- Adjustable effort low · high · max
- Tool calling
- Yes
- Structured output
- Yes
- Weights
- Open weights · MIT
- API model ID
deepseek-v4-proon DeepSeek- Availability
- 37 API providerslisted on models.dev
What DeepSeek claims
Published by the lab at launch. Settings vary, so compare these only with care.
| Benchmark | Score | Setting | Source |
|---|---|---|---|
| Humanity's Last Exam | 42.7 | without tools | Source |
| Humanity's Last Exam | 60 | with tools | Source |
| Terminal-Bench v2.1 | 87.9 | max effort | Source |
| NL2Repo | 61.5 | max effort | Source |
| CyberGym | 83.3 | max effort | Source |
| DeepSWE | 62.7 | max effort | Source |
| Toolathlon-Verified | 74.1 | — | Source |
| Agents' Last Exam | 25.7 | — | Source |
| AutomationBench | 31.8 | — | Source |
| DSBench-FullStack | 71.1 | — | Source |
| DSBench-Hard | 67.2 | — | Source |
More from DeepSeek
About DeepSeek V4 Pro 0813
How much does DeepSeek V4 Pro 0813 cost?
DeepSeek V4 Pro 0813 costs $0.66 input / $1.98 output per million tokens (official DeepSeek API price). Cached input is $0.022 per million tokens. At a 3:1 input-to-output mix that is $0.99 per million tokens, more expensive than 54% of the 360 priced models we track.
What is the context window of DeepSeek V4 Pro 0813?
DeepSeek V4 Pro 0813 accepts up to 1,000,000 tokens per request and can write up to 384,000 tokens in one response.
How good is DeepSeek V4 Pro 0813?
Epoch AI gives DeepSeek V4 Pro 0813 a Capabilities Index score of 155.4 (likely range 153.7–157.6), ranking it #26 of 148 models Epoch has scored. Epoch AI benchmark results: GPQA Diamond 91.7%, FrontierMath Tiers 1–3 64.6%, OTIS Mock AIME 2024–2025 98.6%, SimpleQA Verified 52.9%.
Is DeepSeek V4 Pro 0813 open source?
Yes. DeepSeek publishes the weights under the MIT licence, so it can be downloaded and self-hosted.
What inputs does DeepSeek V4 Pro 0813 support?
DeepSeek V4 Pro 0813 accepts text and replies in text. It is a reasoning model with low, high and max effort settings, supports tool calling and can return structured JSON output.
When was DeepSeek V4 Pro 0813 released?
DeepSeek released DeepSeek V4 Pro 0813 on Aug 12, 2026.
What are the best alternatives to DeepSeek V4 Pro 0813?
The closest current models from other labs on capability, price and release date are Gemini 3.6 Flash (Google, ECI 154.3, $0.75 / $3.75), Muse Spark 1.2 (Meta, ECI 155.0, $1.25 / $4.25), GLM-5.3 (Z.ai (Zhipu), ECI 155.8, $1.40 / $4.40) and GPT-5.6 Luna (OpenAI, ECI 156.5, $0.20 / $1.20).