DeepSeek V4 Flash 0731 is a sparse mixture-of-experts model from DeepSeek, with 13B active parameters out of 284B total. This re-post-trained revision is suited for coding, reasoning, and agent workflows.

Details

FieldValue
iddeepseek/deepseek-v4-flash-0731
familydeepseek-flash
modalitytext->text
context_length1048576
max_output_tokens943718
knowledge_cutoff2025-05
release_date2026-07-31
is_open_weightsyes
supports_tool_callyes
supports_reasoningyes
supports_structured_outputyes
supports_attachmentno
input_modalitiestext
output_modalitiestext
DeepSeek V4 Flash 0731