Skip to content
OmniRail
← All models

DeepSeek · Language model

DeepSeek V4 Flash 0731

deepseek/deepseek-v4-flash-0731
  • Language
Released Jul 31, 2026

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....

Specifications

Context window
1,048,576 tokens
Max output
384,000 tokens
Input
text
Output
text
Reasoning effort
max, high, low, none
Tokenizer
DeepSeek

Supported parameters

Accepted by the model and forwarded as-is by the gateway.

  • frequency_penalty
  • logit_bias
  • logprobs
  • max_tokens
  • min_p
  • parallel_tool_calls
  • presence_penalty
  • reasoning
  • reasoning_effort
  • repetition_penalty
  • response_format
  • seed
  • stop
  • structured_outputs
  • temperature
  • tool_choice
  • tools
  • top_a
  • top_k
  • top_logprobs
  • top_p

Use it

POST /api/v1/chat/completions · MCP chat · full reference

curl https://omnirail.org/api/v1/chat/completions \
  -H "Authorization: Bearer $OMNIRAIL_KEY" -H "Content-Type: application/json" \
  -d '{"model":"deepseek/deepseek-v4-flash-0731","messages":[{"role":"user","content":"Hello"}],"max_tokens":300}'