Neural Inverse is Open Source →
DocsModel Gateway

Model Gateway

Neural Inverse has two separate model services:

ServiceWhat it does
free.neuralinverse.comFree OSS models - DeepSeek, Llama, Mistral, Kimi. No key, free forever.
model.neuralinverse.comPaid third-party models - Claude, GPT-4o, Gemini. Configure provider credentials here.

free.neuralinverse.com

Free forever. No API key, no credit card, no trial expiry. These models show as niFreeModels in the IDE model picker:

ModelNotes
deepseek-r1Reasoning
deepseek-r1-0528Updated reasoning checkpoint
deepseek-v3General purpose
deepseek-v4-flashFast inference
deepseek-v4-proHigher capability
llama-3.3-70bMeta open-weight
llama-4-maverickMeta latest generation
mistral-large-3Mistral open-weight
kimi-k2.6Moonshot AI open-weight

model.neuralinverse.com

Access Claude, GPT-4o, Gemini, and other proprietary models by configuring your provider credentials. Supported providers:

  • Anthropic (Claude)
  • Azure OpenAI
  • AWS Bedrock

Add credentials two ways:

  • In the IDE: Settings, LLM Providers, enter your provider key
  • In the dashboard: Go to model.neuralinverse.com, open Provider Configuration

Once configured, click "Connect to Cloud IDE" in the model dashboard to automatically inject the model service credentials into all your workspaces.

API Keys

model.neuralinverse.com issues API keys for programmatic access. Generate them from the dashboard under API Keys.


Was this page helpful?

Last edited