DeepSeek V4 Pro
DeepSeek V4 Pro is a large-scale reasoning model developed by DeepSeek-AI. It has a gigantic scale with approximately 16 trillion parameters and an extensive context window of 1 million tokens.
Parameters
16000.0B
Context Window
1M
License
MIT
Release Date
2026-04-24
API Pricing
Input Price (per 1M tokens)
$0.435
Output Price (per 1M tokens)
$0.87
Billing Mode: standard
Strengths
- ・Overwhelming 16-trillion parameter count
- ・Long-context understanding of 1M tokens
- ・High openness via MIT license
Weaknesses
- ・Computational load from huge model size
- ・Increased resource consumption during inference
- ・Expected high operational costs
Use Cases
- ・Extremely complex logical reasoning tasks
- ・Analysis of vast documents
- ・Research requiring advanced specialized knowledge
Deep Analysis
Artificial Analysis Intelligence Index
52
#2 open-weights reasoning model, behind Kimi K2.6 (54)
SWE-Bench Verified
80.6%
Ties Claude Opus 4.6 (80.8%) and Gemini 3.1 Pro (80.6%)
LiveCodeBench Pass@1
93.5
#1 among all tested models, ahead of Gemini 3.1 Pro (91.7)
GPQA Diamond
90.1%
Competitive with GPT-5.4 (93.0%) and Gemini 3.1 Pro (94.3%)
Output Price
$3.48/1M (list)
~7x cheaper than Claude Opus 4.7 ($25/1M)
Context Window
1M tokens
8x expansion over V3.2 (128K)
Total Parameters
1.6T (49B active)
MoE architecture; largest open-weights model to date
Strengths
- ・Leads all public models on competitive coding benchmarks (LiveCodeBench 93.5, Codeforces 3206)
- ・MIT-licensed open weights enable self-hosting, fine-tuning, and full compliance control
- ・7-8x cheaper output pricing than Claude Opus 4.7 at near-identical SWE-Bench scores
Weaknesses
- ・Extremely high hallucination rate of 94% — nearly always responds even when uncertain
- ・Political censorship embedded in training weights, not just API-level filtering
- ・Meaningful knowledge gap on SimpleQA (57.9%) vs Gemini 3.1 Pro (75.6%) and HLE (37.7% vs 44.4%)
Competitor Comparison
| Model | SWE | GPQA | Price |
|---|---|---|---|
| Claude Opus 4.7 | 80.9% | 94.2% | $5.00/$25.00 |
| GPT-5.4 | 80.0% | 93.0% | $2.50/$15.00 |
| Gemini 3.1 Pro | 80.6% | 94.3% | $2.00/$12.00 |
Sources
- DeepSeek is back among the leading open weights models with V4 Pro and V4 Flash
- DeepSeek: DeepSeek V4 Pro - AI Model Details & Benchmarks
- DeepSeek Benchmarks 2026: V4-Pro & V4-Flash Results
- deepseek-ai/DeepSeek-V4-Pro · Hugging Face
- DeepSeek v4: the most expected open-source model ever released, and the quietest landing
- DeepSeek V4 vs Claude Opus 4.7: Pricing, Benchmarks, and Tradeoffs
- DeepSeek V4 Review: 1M Context, Pro vs Flash, Coding, Agents, and Limits
- DeepSeek V4-Pro Review: Frontier Power, Penny Prices
- DeepSeek V4 vs Claude Opus 4.5 for coding: benchmark comparison
Analysis generated: 2026-05-23