D
DeepSeek: deepseek-v3
deepseek-v3
128K context8.2K outReasoningToolsFilesCacheStructured
Released Dec 26, 2024Knowledge cutoff Sep 2025Updated Aug 27, 2026
A state-of-the-art 671B MoE model (37B active) featuring Multi-head Latent Attention (MLA) and auxiliary-loss-free load balancing. It provides a hybrid "Think/Non-think" mode, delivering SOTA performance in coding, math, and agentic tasks while remaining the most cost-efficient flagship in the 2026 market.
Pricing
Input price
$0.56/ 1M tokens51% off
Output price
$2.24/ 1M tokens51% off
Context window 163.8K tokensCompatible endpoints openaiVendor DeepSeek
Uptime
Performance
Loading performance data...
Usage & Ranking
Loading usage...
Supported parameters
All providers = every upstream serving this model supports it. Some providers = depends on which upstream handles the request. Default = the value sent when you leave the parameter unset.
| Parameter | Providers | Default |
|---|---|---|
| frequency_penalty | All providers | Not sent by default |
| include_reasoning | All providers | - |
| logit_bias | Some providers | - |
| logprobs | All providers | - |
| max_tokens | All providers | - |
| min_p | Some providers | - |
| presence_penalty | All providers | Not sent by default |
| reasoning | All providers | - |
| repetition_penalty | Some providers | Not sent by default |
| response_format | All providers | - |
| seed | All providers | - |
| stop | All providers | - |
| structured_outputs | All providers | - |
| temperature | All providers | 1 |
| tool_choice | All providers | - |
| tools | All providers | - |
| top_k | All providers | Not sent by default |
| top_logprobs | All providers | - |
| top_p | All providers | 0.95 |