DeepSeek
DeepSeek is the Chinese AI lab that keeps closing the gap with frontier proprietary models while releasing everything under the MIT license. Their V4 family, launched in April 2026, includes V4 Pro (1.6 trillion parameters, 49B active) and V4 Flash (284B total, 13B active), both with native 1M token context windows. V4 Pro scored 80.6% on SWE-bench Verified, within 0.2 points of Claude Opus 4.6. At $1.74 per million input tokens for Pro and $0.14 for Flash, they offer near-frontier performance at a fraction of proprietary pricing.
Service status
Loading…
Founded
2023
Headquarters
Hangzhou, China
CEO
Liang Wenfeng
Models
2 active
Key Products
Strengths
- ✓MIT open source license
- ✓Near-frontier benchmarks
- ✓Ultra-competitive pricing
- ✓Native 1M context
- ✓Strong coding performance
DeepSeek Models
| Model | Input / 1M | Output / 1M | Context | Capabilities |
|---|---|---|---|---|
| DeepSeek V4 Pro | 0.43 | 0.87 | 1M | text, vision, code, reasoning |
| DeepSeek V4 Flash | 0.14 | 0.28 | 1M | text, vision, code |
Prices per 1M tokens in USD. See the full pricing guide for detailed analysis.
Benchmark Scores
| Model | SWE-bench | MMLU-Pro | HumanEval | GPQA Diamond | MATH |
|---|---|---|---|---|---|
| DeepSeek V4 Pro | 80.6 | 91.5 | 94.8 | 73.1 | 92.4 |
| DeepSeek V4 Flash | 79.0 | 85.2 | 89.4 | 58.7 | 82.1 |
| DeepSeek V3 | 42.0 | 88.1 | 91.2 | 63.5 | 85.9 |
Comparisons
DeepSeek V4 Pro vs GPT-5.5
DeepSeek vs OpenAI
Claude Opus 4.7 vs DeepSeek V4 Pro
Anthropic vs DeepSeek
Llama 4 Maverick vs DeepSeek V4 Pro
Meta vs DeepSeek
DeepSeek V4 Flash vs Gemini 2.0 Flash
DeepSeek vs Google
LongCat-2.0 vs DeepSeek V4 Pro
Meituan vs DeepSeek
MiniMax M3 vs DeepSeek V4 Pro
MiniMax vs DeepSeek