Product Guide
Nexum Router
The complete guide to Nexum Router: what it costs, what that price actually buys, which models are included, how the endpoints work, and how the usage comparison against Claude Code Max is measured.
Last updated: 2026-08-31
What Nexum Router is
Nexum Router is a hosted LLM gateway. You subscribe once, create an API key in the dashboard, and point your existing tools at one stable endpoint. There are no hourly limits, no weekly caps, and no five-hour session windows — one flat $5.50/week plan.
- One flat $5.50/week — every model and every tool included
- Native Claude Code endpoint (set ANTHROPIC_BASE_URL and go)
- Flat $5.50/week — cancel anytime from the dashboard
- Live status page and a real usage dashboard
Endpoints and model IDs
/router/v1 as the base URL; Claude Code uses /router/claude. Switch models by changing the model ID — nothing else in your setup changes.- Qwen 3.7 Max — qwen-3.7-max (thinking variant: qwen-3.7-max-thinking)
- Xiaomi Mimo 2.5 — xiaomi-mimo-2.5
- Meta Muse Spark 1.2 — meta-muse-spark-1.2
- Meta Muse Spark 1.3 — meta-muse-spark-1.3
How the usage comparison is measured
We compare the weekly token throughput the plan sustains against the published usage limits of Claude Code Max, which enforces five-hour rolling windows plus weekly caps. The plan matches or beats the ceiling of the top $200/month Max 20x tier, for $5.50/week — roughly one-eighth the price. Nexum Router has no windows and no caps, which is where most of the difference comes from. The live leaderboard on the homepage shows real token volume flowing through the router.
- Same-or-better ceiling than the $200/month Max 20x tier
- $5.50/week — about $24 a month
- No five-hour session windows, no weekly usage caps
- Real usage visible on the public leaderboard
Client compatibility
Anything that speaks the OpenAI API works: the official SDKs, curl, opencode, OpenClaw, Hermes, Cline, and similar tools. Claude Code is supported through the dedicated Anthropic-compatible endpoint. Xiaomi Mimo 2.5 covers everyday work, and Qwen 3.7 Max handles flagship coding and agentic sessions.
- OpenAI SDK, curl, and any OpenAI-compatible client
- Claude Code ready out of the box
- opencode, OpenClaw, Hermes, and other agent CLIs