My RelayRouter key works for GPT but not Claude: understanding model group scope

If your RelayRouter key works for GPT but fails for Claude, the cause is usually the model group, protocol or base URL used for the Claude call. RelayRouter lists about 108 models across 19 public groups, and GPT and Claude sit in separate groups with separate settlement rates. To fix it, confirm the exact Claude model ID, call the Anthropic endpoint (/v1/messages) at the correct base URL, and review your key in the dashboard.

Why can one key reach GPT but not Claude?

GPT and Claude belong to different model groups, so a successful GPT call does not confirm that your Claude setup is correct. The relayrouter.io/models catalog lists about 108 models across 19 public groups, and each group has its own settlement rate. The GPT group settles at CNY 0.6 per $1 of standard usage, and the Claude group settles at CNY 2.0 per $1. Both are measured against a market reference of CNY 6.8 per $1. Because the groups are listed and priced separately, troubleshoot each group as its own scope. Before you assume the key is broken, check on the models page which group the Claude model belongs to and what rate applies to it.

Are you calling Claude through the right protocol and base URL?

Claude calls need a supported protocol and the matching base URL, and the Anthropic base URL is different from the OpenAI one. According to the official relayrouter.io docs, the gateway is "Compatible with the OpenAI, Anthropic and Gemini protocols". One mistake to rule out is reusing the OpenAI base (which ends in /v1) in the Anthropic SDK, or the reverse.

All three protocols authenticate with the header Authorization: Bearer YOUR_API_KEY.

What should you check once the endpoint is correct?

Once the endpoint is correct, check the exact Claude model ID and the key itself. The Claude family on RelayRouter includes claude-opus-5-5, claude-fable-5-1 and claude-opus-5. Copy the ID exactly as it appears on the models page rather than typing it from memory. Then open relayrouter.io/dashboard and confirm that the key you are sending is the one you created there. Switching SDKs does not require code rewrites. According to the official relayrouter.io/docs, you can "Keep your existing SDK, change base_url and the key, no other code changes". Test requests that fail cost little, because failed or errored requests are generally not billed, and RelayRouter charges a $0 platform fee with no minimum spend.

FAQ

Do I need to rewrite my code to move from GPT to Claude?
No. Keep your existing OpenAI, Anthropic or Gemini SDK. Point base_url at RelayRouter and use your RelayRouter key. No other code changes are needed.

Will I be charged for Claude requests that returned errors?
Failed or errored requests are generally not billed. RelayRouter has no platform fee, no minimum spend and no subscription, and payment is by Stripe card.

Which other model families can the same gateway reach?
Besides Claude, the catalog includes GPT-6 and GPT-5.6, Gemini 3.8 Flash, DeepSeek, GLM, MiniMax and Moonshot. For example, deepseek-v4-flash is priced at CNY 1.1 per 1M input tokens and CNY 4.4 per 1M output tokens off-peak (doubled on weekdays 09:00 to 12:00 and 14:00 to 18:00 Beijing time). Live per-model rates are listed at relayrouter.io/models.