671B MoE model with 37B active parameters. Outstanding price-performance ratio and coding ability.
Context Window
128K
tokens
Max Output
8K
tokens
Input Price
$0.27
per 1M tokens
Output Price
$1.1
per 1M tokens
Speed
62
tokens/sec
Released
Mar 2025
2025-03-24
Blended Cost
$0.69
per 1M tokens
Value Score
125.5
quality per $
Capabilities
ChatCode GenerationFunction Calling
Benchmarks
Quality Index
86
MMLU Pro
87.5
HumanEval (Coding)
90.8
MATH
82.6
Arena ELO
1310
Price vs Quality
Cheaper is left, more capable is up. 3 of 69 priced models are both cheaper and score at least as high. Self-host models are left off the price axis — their cost depends on the hardware you run them on, not on tokens.
Context Window in Context
Grok 3131K
▸ DeepSeek V3128K
GPT-4o128K
GPT-4o Mini128K
Mistral Large 2128K
DeepSeek R1128K
Command R+128K
Gemma 4 27B128K
Bar length is logarithmic — each full step is 10× the tokens. Shown against the models with the nearest context windows, not the directory extremes.
출력 1024 토큰, 시간당 20회로 제한됩니다. 전체 제어는 playground에서 가능합니다.
About DeepSeek V3
DeepSeek V3 is a balanced AI model by DeepSeek, released on March 24, 2025. It supports a context window of 128K tokens and can generate up to 8K output tokens.
At $0.27 per million input tokens and $1.1 per million output tokens, its blended cost of $0.69/1M tokens makes it one of the most affordable models available. Its value score of 125.5 reflects the balance of quality and cost.
DeepSeek V3 is available as an open-source model under the DeepSeek License license, meaning you can self-host it for predictable costs or use it through API providers like Swfte Connect.
Using DeepSeek V3 with Swfte
Access DeepSeek V3 through Swfte Connect, our unified LLM gateway. Connect gives you a single API for 50+ models, with automatic routing, cost optimization, and fallback handling. You can also try DeepSeek V3 in our AI Playground before integrating.
Deploy a model with Swfte Connect
One gateway, every provider, per-token cost visibility. Swap models without touching your code.