Router9
deepseek/deepseek-v4-flash
Context window1.0M tokens
Free budget10,000 credits/month (~$1), shared; account-anniversary reset tokens/mo
Rate limitsnot published
Shared monthly free credits info
Router9 Free grants 10,000 credits (about $1) per monthly billing period, shared across API keys and Studio usage. The allowance renews on the account signup anniversary, not necessarily the first of the month, and unused credits do not roll over. Paid prepaid overage is outside the free allowance. Verified against current provider documentation November 3, 2026. https://www.router9.com/docs/token-plan/credits
Unversioned route resolves to 0731 info
On September 14, 2026, deepseek/deepseek-v4-flash returned a valid completion identifying deepseek/deepseek-v4-flash-0731. The public free-model listing still labels this route 0423. The catalog displays the observed 0731 revision while preserving the tested request ID. Response identifiers are provider assertions, not independent verification of weights.
ElectronHub
deepseek-v4-flash-0731
Context window524K tokens
Free budget$0.25/week shared free credits tokens/mo
Rate limitsnot published
Shared weekly free credits info
ElectronHub Free renews $0.25 each week (Sunday 21:00 UTC), shared across eligible non-premium models; unused weekly credit does not roll over. These are credit-backed routes, not the separate daily :free aliases. Paid top-ups are optional and not included in this allowance. Text inference was tested; other capability flags come from the provider roster.
Inferbase
deepseek-v4-flash-0731
Context window8K tokens
Free budget5,000 requests/month; 100 requests and $0.50 usage/day shared tokens/mo
Rate limitsnot published
Shared monthly and daily free allowance info
Inferbase Free needs no card. The shared account allowance is 5,000 requests/month, also capped at 100 managed requests and $0.50 of model usage/day. Free managed chat permits 8,192 input and 1,024 output tokens per request. Counts are shared, not per-model grants. Model IDs including FP8 and Turbo variants are explicit; a changed response identity fails rather than silently substituting another model. Update FreeLLMAPI to a build with the Inferbase adapter. See https://inferbase.ai/pricing.
Get this model the moment it changes
Limits move, models get replaced, better ones launch. Premium routers see this page's data live. Free routers see last month's.
লাইভ করুন · $19/yr →Use it
FreeLLMAPI is a self-hosted router you run yourself. Install it, paste in your free Router9 key, and deepseek/deepseek-v4-flash answers on an OpenAI-compatible endpoint at http://localhost:3001/v1. No credit card, no hosted middleman: your prompts and your provider keys never leave your machine.
Install the router (macOS, Linux, WSL)
curl -fsSL https://freellmapi.co/install.sh | bashInstall the router (Windows PowerShell)
iwr -useb https://freellmapi.co/install.ps1 | iexCall deepseek/deepseek-v4-flash with curl
curl http://localhost:3001/v1/chat/completions \
-H "Authorization: Bearer freellmapi-your-unified-key" \
-H "Content-Type: application/json" \
-d '{
"model": "deepseek/deepseek-v4-flash",
"messages": [{"role": "user", "content": "Say hi in five words."}]
}'The same request in Python (openai)
from openai import OpenAI
client = OpenAI(
base_url="http://localhost:3001/v1",
api_key="freellmapi-your-unified-key",
)
resp = client.chat.completions.create(
model="deepseek/deepseek-v4-flash",
messages=[{"role": "user", "content": "Say hi in five words."}],
)
print(resp.choices[0].message.content)The router answers on /v1/chat/completions and every other OpenAI surface, plus the Anthropic Messages API, so existing clients need only a new base_url. Swap the model id for auto and the router picks the best free model that is still under its limits. Full reference: docs/api.md.