AnyModelAnyModel
Get started
DeepSeek

DeepSeek V4 Flash

The cheapest serious reasoner on the market.

DeepSeek V4 Flash is the volume tier of the V4 generation — the model the old deepseek-chat/reasoner aliases now route to, with a 1M-token window at prices that make bulk reasoning practically free.

Best for

Bulk jobsBudget reasoningHigh volume

Strengths

  • ▸Rock-bottom price
  • ▸1M-token context
  • ▸Solid reasoning

Use DeepSeek V4 Flash via AnyModel

One OpenAI-compatible endpoint — just set the model id:

curl https://anymodel.org/v1/chat/completions \
  -H "Authorization: Bearer $ANYMODEL_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "am/deepseek-v4-flash",
    "messages": [{"role": "user", "content": "Hello!"}]
  }'
One-line Codex installGet a key