ClawCloud gives managed OpenClaw instances a curated model catalog that is ready on day one. Pick a plan, connect your channel, then switch models in chat with /modelinstead of rebuilding config every time.
Curated defaults across frontier, budget, and coding-focused models in managed mode.
OpenAI, Anthropic, Google, DeepSeek, Mistral, Llama, Grok, Qwen, and more.
Optional Backup Models addon enables supported :free models as a manual recovery path.
Managed mode is the fastest path for most buyers. ClawCloud provisions the API key, applies the curated model list, tracks usage, and lets you move between supported models in chat without touching provider dashboards.
GPT-4.1, GPT-5.3 Codex, o3, o4-mini
Claude Haiku 4.5, Sonnet 4.6, Opus 4.6
Gemini 2.5 Flash, Flash Lite, 3.1 Pro Preview
Grok 3, Grok 4, Grok Code Fast 1
Llama 3.1, Llama 3.3, Llama 4 Maverick, Scout
Codestral, Devstral, Mistral Large, Mistral Medium
Qwen3 coder, max, 235B
DeepSeek Chat V3.1, R1, V3.2
Nova Premier, Nova Pro
Mercury, Mercury Coder
Seed 1.6
Command R+
Kimi K2.5
Llama 3.1 Nemotron 70B
Bring your own Anthropic, OpenAI, or Google key if you want direct provider billing. You keep the same dedicated server and dashboard, but your available models come from the provider tied to that key.
Use your Anthropic key for Claude Haiku, Sonnet 4.6, and Opus variants.
Use your OpenAI key for GPT-4.1, GPT-5.x, o-series, and Codex models.
Use your Google key for Gemini 2.5 and Gemini 3.1 preview models.
Backup Models is optional. It does not switch automatically. When credits run low, it gives you supported :freemodels you can move to manually in chat so the bot can keep responding.
The sales advantage here is speed: customers do not need to redeploy just to test a different model family. Managed mode keeps the catalog ready and switching happens in chat. That lowers friction for evaluation, cost control, and recovery when a cheaper or free option is good enough.