Comparison

Nexum Router vs OpenRouter for coding agents

OpenRouter charges each model at its provider's list price and adds a 5.5% fee when you buy credits. Nexum Router charges a flat $5.50 a week. Which is cheaper depends on how many tokens your agent burns in a day.

By the Dialagram team Updated 2 min read

How OpenRouter charges

  • No markup on inference: every plan bills each model at its provider's list price
  • A platform fee when you buy credits: 5.5% on Standard, 8% on Business
  • Bring your own key: the first $25,000 of list-price inference a month carries no fee, then 5%
  • Free models: 20 requests a minute and 50 a day, rising to 1,000 a day once you have bought at least $10 of credits; failed attempts count toward the cap

Side by side

OpenRouter, StandardNexum Router
PriceProvider list price per token, plus 5.5% on credit purchases$5.50/week flat (about $22 a month)
Models500+ models from 80+ providers20 models: DeepSeek, Qwen, Xiaomi and Meta
MeterEvery tokenNo hourly or weekly windows on Qwen, Xiaomi and Meta; $5 of DeepSeek V4.1 Flash a day included
ClientsOpenAI-compatible APIOpenAI-compatible API plus a native Claude Code endpoint

The break-even point

$5.50 a week is about $0.79 a day. If your agent's per-token bill would stay under that, pay-per-token is cheaper. Agents rarely stay there: they resend the conversation on every turn, so a day of steady coding moves millions of input tokens. At DeepSeek V4.1 Flash's list price of $0.30 per million uncached input tokens, about 2.6 million uncached tokens already cost $0.79, before output and before any pricier model.

When OpenRouter is the better pick

  • You need Claude, GPT, Gemini or another model Nexum Router does not serve
  • You use models occasionally, so a per-token bill stays small
  • You already hold provider keys and want one API in front of them
  • You want free models for experiments and can live with 50 requests a day

When a flat plan is the better pick

  • Your coding agent runs for hours every day
  • You want a bill that does not move with token count
  • You use Claude Code and want a native Anthropic-compatible endpoint

Switching takes one setting

Both speak the OpenAI API, so moving a tool between them means changing its base URL and key. For Nexum Router the base URL is /router/v1, and Claude Code uses /router/claude; see the Claude Code setup guide.

Frequently asked questions

Does OpenRouter mark up model prices?

No. OpenRouter bills inference at the provider's list price and charges its fee, 5.5% on the Standard plan, when you buy credits.

What are OpenRouter's free model limits?

20 requests a minute and 50 a day, rising to 1,000 a day after you have bought at least $10 of credits.

Can I use OpenRouter and Nexum Router together?

Yes. Each tool takes its own base URL, so you can point different tools, or different model choices, at each service.

Sources

Prices, limits and features were checked against these pages on 2026-10-29.