OpenRouter
OpenRouter provides access to 300+ AI models from all major providers through a single API key. This includes many free models that require no payment -- just a free account.
Getting a Free API Key
- Go to openrouter.ai/keys
- Create a free account (no credit card required)
- Generate an API key (starts with
sk-or-...) - Paste it in AI Supreme Council under Settings > AI Model > OpenRouter
OpenRouter is the best way to access models from multiple providers without managing separate API keys. One key gives you access to Claude, GPT, Grok, Gemini, Llama, DeepSeek, Qwen, and many more.
Free Models
OpenRouter offers free access to many open-source and community models. These require an API key but have no per-token cost.
Featured Free Models
| Model | Developer | Context | Strengths |
|---|---|---|---|
| DeepSeek R1 | DeepSeek | 164K | Reasoning, chain-of-thought |
| Qwen 3 Coder 480B | Alibaba | 262K | Code generation, massive model |
| Qwen 3 235B | Alibaba | 131K | General purpose, large model |
| Qwen 3 Next 80B | Alibaba | 262K | Next-gen compact model |
| GLM 4.5 Air | Zhipu AI | 131K | General purpose, tools |
| Llama 3.3 70B | Meta | 128K | General purpose, tools |
| GPT-OSS 120B | OpenAI (open) | 131K | General purpose, reasoning |
| Aurora Alpha | OpenRouter | 128K | Reasoning, tools |
| Pony Alpha | OpenRouter | 200K | Reasoning, large output |
| Mistral Small 3.1 | Mistral | 128K | Vision, efficient, tools |
| Gemma 3 27B | 131K | Vision, compact, tools | |
| Step 3.5 Flash | StepFun | 256K | Large context, fast |
| Trinity Large 400B | Arcee AI | 131K | Large model, tools |
Free models on OpenRouter have rate limits:
- Without credits: ~50 requests/day
- With $5+ credits: ~1,000 requests/day
Adding even a small amount of credits to your account significantly increases your rate limit, even when using free models.
Paid Models
OpenRouter also provides access to all major paid models through a single API key:
| Provider | Available Models |
|---|---|
| Anthropic | Claude Opus 4.6, Opus 4.5, Sonnet 4.5, Sonnet 4, Opus 4 |
| OpenAI | GPT-5, GPT-4o, o3, o3 Mini |
| xAI | Grok 4.1 Fast, Grok 4 Fast, Grok 4, Grok 3 |
| Gemini 3 Pro/Flash Preview, Gemini 2.5 Pro, Gemini 2.5 Flash | |
| Meta | Llama 4 Scout, Llama 4 Maverick |
| Mistral | Mistral Large, Codestral |
| DeepSeek | DeepSeek R1, V3 |
Pricing on paid models matches or is close to the original provider's rates. Using OpenRouter for paid models gives you the convenience of a single API key and single billing account across all providers.
How to Select Models
When you choose OpenRouter as the provider in a bot profile, the model dropdown shows all available OpenRouter models from the community registry. Models are identified by their full path (e.g., qwen/qwen3-coder:free, openrouter/aurora-alpha).
Free models typically have :free in their model ID or (free) in their display name.
Pricing
- Free models: No cost per token. Rate limited.
- Paid models: Pay per token, billed to your OpenRouter account. Prices vary by model and are displayed on openrouter.ai/models.
- Credits: Prepay credits via credit card or crypto. Unused credits do not expire.
AI Supreme Council tracks token usage per provider in Settings > Usage.
Features
OpenRouter supports all major features through its OpenAI-compatible API:
| Feature | Supported |
|---|---|
| Streaming (SSE) | Yes |
| Vision (image input) | Yes (model-dependent) |
| Tool/function calling | Yes (model-dependent) |
| Reasoning | Yes (model-dependent) |
| JSON mode | Yes (model-dependent) |
Feature availability depends on the underlying model. For example, vision works with GPT-4o and Gemini models but not with text-only models like DeepSeek R1.
Configuration
When creating a bot profile, select OpenRouter as the provider and choose your preferred model. You can set a per-bot API key in the bot configuration panel to override the global key.
The OpenRouter provider uses the OpenAI-compatible Chat Completions API (openrouter.ai/api/v1/chat/completions) with additional headers (HTTP-Referer and X-Title) for attribution.
Tips for Best Results
- Start with free models. DeepSeek R1 (reasoning), Qwen 3 Coder (code), and Llama 3.3 70B (general) are excellent free options.
- Add $5 credits for better rate limits. Even if you only use free models, having credits on your account increases your daily request limit from ~50 to ~1,000.
- Use OpenRouter for council mode. When building multi-model councils, OpenRouter lets you mix models from different providers without managing multiple API keys.
- Check model availability. OpenRouter's model catalog changes frequently. New models are added regularly, and some may be temporarily unavailable.