Skip to content
i want to see how is the pricing for agents when used and how to change model in the agents chat?
jwtiyar
OP

5 days ago

I want to know the pricing for agents in raiwlay agents chat, also before new ui updates for railway there was option to change model but now i cant find it.

Awaiting User Response

7 Replies

Railway
BOT

5 days ago

The Railway Agent is billed on the LLM tokens it uses, at the model provider's per-token rates with no markup, and it draws from the same included credit as your compute (the $5 on Hobby). You're only charged extra once combined compute and agent spend goes past that, and agent usage shows as its own line on your invoice. Per-model token rates are on anthropic.com/pricing, and the details are in our pricing docs.

When the agent runs on our models, it routes each request automatically to the smallest model that can handle it (Claude Haiku 4.5, Sonnet 4.6, or Opus 4.6). If you want to run it on a model account of your own, Bring Your Own Models lets you sign in with ChatGPT or OpenRouter, and those requests are billed by that provider instead.

You can see your agent spend and set an agent hard limit (down to $0) from your workspace usage page via "Update limits".


Status changed to Awaiting User Response Railway • 5 days ago


jwtiyar
OP

5 days ago

where do i change the agents in the chat?


Status changed to Awaiting Railway Response Railway • 5 days ago


Railway
BOT

5 days ago

Agent chat has a model picker that lets you switch the model for a single message. Your default model is set under Default Model in Account Settings, then Agents, with the right workspace selected.

To add models from your own ChatGPT or OpenRouter account, first turn on the matching flag on the Feature Flags page. Then connect the provider in the Providers section of Agents settings. Details are in Bring Your Own Models.


Status changed to Awaiting User Response Railway • 5 days ago


Railway

Agent chat has a model picker that lets you switch the model for a single message. Your default model is set under **Default Model** in [Account Settings, then Agents](https://railway.com/account/agents), with the right workspace selected. To add models from your own ChatGPT or OpenRouter account, first turn on the matching flag on the [Feature Flags page](https://railway.com/account/feature-flags). Then connect the provider in the **Providers** section of Agents settings. Details are in [Bring Your Own Models](https://docs.railway.com/ai/bring-your-own-models).

jwtiyar
OP

5 days ago

agent chat no longer has model picker


Status changed to Awaiting Railway Response Railway • 5 days ago


Railway
BOT

5 days ago

You can still change the model without the chat picker. Go to Account Settings, then Agents with the right workspace selected, and pick a model under Default Model. Railway Agent chat uses that default for every message.

Since the picker isn't showing in your chat, could you attach a screenshot of the chat input area where you'd expect to see it?


Status changed to Awaiting User Response Railway • 5 days ago


jwtiyar
OP

5 days ago

screenshot-2026-11-14_21-18-39-firefox.png


Status changed to Awaiting Railway Response Railway • 5 days ago


Railway
BOT

5 days ago

Your screenshot shows a cloud agent conversation, not the dashboard's Railway Agent chat. The per-message model picker is only in the dashboard chat. Cloud agents you own use the model set under Default Model in Account Settings, then Agents, with that workspace selected.

The model is locked when a conversation starts. To use a different one, change Default Model and then start a new agent with "New agent". Your existing conversations keep their current model.


Status changed to Awaiting User Response Railway • 5 days ago


Welcome!

Sign in to your Railway account to join the conversation.

Loading...