DeepSeek V4 Pro
Open MoE flagship with million-token context for coding and long agent runs.
- Capability
- 150.2ECI · #67 of 158
- Input
- $1.60per 1M tokens
- Output
- $3.33per 1M tokens
- Context
- 1M384K max output
What it costs
- Input
- $1.60per million tokens
- Output
- $3.33per million tokens
- Blended (3:1)
- $2.03pricier than 73% of priced models
Typical monthly bills
| Monthly usage | Estimated cost |
|---|---|
| Side project2M input + 0.5M output tokens | $4.87 |
| Team assistant25M input + 5M output tokens | $56.65 |
| Production app250M input + 50M output tokens | $566.50 |
Median of 65 paid API providers on models.dev.
Independent benchmarks
#67 of 158 scored models
The shaded band is Epoch AI’s confidence range (148.5–152.0); the tick marks the median scored model.
Scores from Epoch AI, run independently of DeepSeek.
The details
- Lab
- DeepSeek
- Released
- Apr 24, 2026
- Knowledge cutoff
- May 2025
- Context window
- 1,000,000 tokens
- Max output
- 384,000 tokens
- Inputs
- Text
- Output
- Text
- Reasoning
- Always on
- Tool calling
- Yes
- Structured output
- No
- Weights
- Open weights
- Availability
- 67 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 |
|---|---|---|---|
| SWE-Bench Verified | 80.6 resolved | — | Source |
| Artificial Analysis Coding Agent Index | 50.1 | high | Source |
| SWE-Atlas Codebase QnA | 67.8 | high | Source |
| SWE-Bench Pro | 18 | high | Source |
| Terminal-Bench v2.1 | 64.7 | high | Source |
| MMLU-Pro | 87.5 EM | preview checkpoint; max effort | Source |
| SimpleQA-Verified | 57.9 | preview checkpoint; max effort | Source |
| Chinese SimpleQA | 84.4 | preview checkpoint; max effort | Source |
| GPQA Diamond | 90.1 | preview checkpoint; max effort | Source |
| Humanity's Last Exam | 37.7 | preview checkpoint; max effort; without tools | Source |
| LiveCodeBench | 93.5 | preview checkpoint; max effort | Source |
| Codeforces | 3206 rating | preview checkpoint; max effort | Source |
More from DeepSeek
About DeepSeek V4 Pro
How much does DeepSeek V4 Pro cost?
DeepSeek V4 Pro costs $1.60 input / $3.33 output per million tokens (median across 65 API providers). At a 3:1 input-to-output mix that is $2.03 per million tokens, more expensive than 73% of the 469 priced models we track.
What is the context window of DeepSeek V4 Pro?
DeepSeek V4 Pro 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?
Epoch AI gives DeepSeek V4 Pro a Capabilities Index score of 150.2 (likely range 148.5–152.0), ranking it #67 of 158 models Epoch has scored. Epoch AI benchmark results: GPQA Diamond 90.9%, FrontierMath Tiers 1–3 45.3%, OTIS Mock AIME 2024–2025 96.7%, SWE-bench Verified 77.6%, SimpleQA Verified 47.0%.
Is DeepSeek V4 Pro open source?
Yes. DeepSeek publishes the weights, so it can be downloaded and self-hosted.
What inputs does DeepSeek V4 Pro support?
DeepSeek V4 Pro accepts text and replies in text. It is a reasoning model, supports tool calling.
When was DeepSeek V4 Pro released?
DeepSeek released DeepSeek V4 Pro on Apr 24, 2026. Its training data runs to May 2025.
What are the best alternatives to DeepSeek V4 Pro?
The closest current models from other labs on capability, price and release date are GLM-5.1 (Z.ai (Zhipu), ECI 150.7, $1.40 / $4.40), Grok 4.3 (xAI, ECI 149.9, $1.25 / $2.50), Kimi K2.6 (Moonshot AI, ECI 151.6, $0.95 / $4.00) and GPT-5.4 mini (OpenAI, ECI 149.4, $0.75 / $4.50).