Model page

DeepSeek V4 Flash

DeepSeek-V4-Flash via Fireworks: streamlined open-source MoE model optimized for fast, cost-efficient inference while preserving strong reasoning and coding performance at 1M context scale. Function calling supported. Uses 1 base request per send before length multipliers. Does not support web search or image input.

Key dates

We use concrete dates and keep them consistent across the site.

Capability

Intelligence

High

Capability

Speed

Fast

Capability

Context window

1,000,000 tokens

Modalities

Input and output

Input: Text
Output: Text

Features

Availability notes

Cached input: $0.03 / 1M tokens · 1 base request per send before length multipliers · Function calling supported · Serverless through Fireworks

ChatGPT

GPT‑4o retirement (ChatGPT)

OpenAI is retiring GPT‑4o inside ChatGPT on February 13, 2026. This does not automatically imply API removal.