Models
The 18-model catalog available with a LEVEL11 account, the auto router, which models each plan may use, and bring-your-own-key (BYOK).
The account catalog: seven providers, 18 models
Signed in with a LEVEL11 account, every catalog model runs through the same gateway. Provider keys stay on the server; the app holds only a device token.
| Provider | Models |
|---|---|
| Anthropic | Claude Fable 5.1, Claude Fable 5, Claude Opus 5.5, Claude Haiku 5.5 |
| OpenAI | GPT-5.6 Sol, GPT-5.6 Luna, GPT-6 Astra, GPT-6.1 Sol, GPT-6 Luna |
| Gemini 4 Argon, Gemini 3.8 Flash | |
| xAI | Grok 4.7 |
| Alibaba | Qwen 3.8 Max |
| Workers AI (open-weight models hosted by Cloudflare) | DeepSeek V4.1 Flash, DeepSeek V4 Pro, GLM 5.3, GLM 5.3 Flash, Qwen 3.8 27B |
Claude is called through the Anthropic Messages API, so adaptive thinking (effort) and prompt caching work unchanged. The GPT-6 family uses the OpenAI Responses API; everything else uses OpenAI Chat Completions. Thinking depth maps to effort for Claude, reasoning effort for GPT and a thinking budget for Gemini. Change the model any time from the composer or Settings → Model; mid-session switches keep the conversation and tool calls.
auto: the LEVEL11 Router
With the model set to auto, the gateway picks a model per request for your plan:
- requests with tools attached (agent work) → the plan's agentic model
- short requests with
max_tokensof 1,024 or less → the light model - everything else → the default model
On Free all three are Claude Haiku 5.5. On Pro and Team the default and agentic model is Claude Opus 5.5 and the light model is Claude Haiku 5.5. The x-level11-model response header and the turn details in the panel show which model actually answered.
Models per plan
- Free: the value line of seven — Claude Haiku 5.5, GPT-6 Luna, GPT-5.6 Luna, Gemini 3.8 Flash, DeepSeek V4.1 Flash, GLM 5.3 Flash, Qwen 3.8 27B. Output is capped at 16K tokens per request and the request body at 600 KB (roughly 150K tokens of context).
- Pro and Team: all 18. Output up to 64K (Pro) or 128K (Team) tokens, request body up to 4 MB, so the full 1M-token context window is usable.
Picking a model that is not on your plan is refused by the gateway with the list of models the plan allows. Limits and prices are in Account and billing.
Your own API keys (BYOK)
Choose Use my own API key under Settings → LEVEL11 account to connect directly to Anthropic, OpenAI and Google Gemini with your keys. Pick from the built-in list or type a custom model ID. Before you sign in, Claude models use the Anthropic key (or ANTHROPIC_API_KEY); once signed in, every catalog model runs through the account and the stored keys are used only in BYOK mode. BYOK usage is not metered by LEVEL11; it is billed by each provider to your own account.
xAI, Alibaba, DeepSeek and Z.ai models are not BYOK providers; they run only through a LEVEL11 account. If one of them is selected while signed out, the app asks you to sign in from Settings.
Local models
The fourth option is a model server on your own machine. Choose Local model in the picker to connect to Ollama, llama.cpp or MLX directly; nothing is metered. See Local models.