Hub/Models/deepseek/deepseek-v4-flash-0731
deepseektext → text

DeepSeek: DeepSeek V4 Flash 0731

Canonical ID: deepseek/deepseek-v4-flash-0731

Compare ModelOpen on OpenRouter →

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....

Artificial Analysis Benchmarks

Verified
No standardized Artificial Analysis benchmark scores recorded for this variant yet.

Token Pricing & Economics

Prompt Input (1M Tokens)
$0.04
Input
Completion Output (1M Tokens)
$0.30
Output
Prompt KV-Cache Read (1M Tokens)
$0.04
Cached
Context Window Capacity
1,024k tokens
Context

Supported Parameters & Features

frequency_penaltyinclude_reasoninglogit_biaslogprobsmax_tokensmin_pparallel_tool_callspresence_penaltyreasoningreasoning_effortrepetition_penaltyresponse_formatseedstopstructured_outputstemperaturetool_choicetoolstop_atop_ktop_logprobstop_p