Custom API Keys fail with "The model does not work with your current plan or api key"

Bug Reports
1

I’ve been using my own OpenRouter API key for the last months without any problem. Suddenly, since last week I’m getting the error the model does not work with your current plan or api key. It happens with some models (gemini, llama, deepseek). I’ve tried with different namings (google/gemini-2.5-pro-preview, gemini-2.5-pro-preview) but still doesn’t work.

Weirdly gpt-4o works but openai/gpt-4o fails.

I’m on a free plan but using custom API keys. Without this feature it doesn’t make sense for me to use Cursor.

10 Likes
2

me too

3

Hey, thanks for the report. We’ll try to fix it.

1 Like
4

yeah same issue. Very annoying. Have tried a bunch of options/workarounds at this point none work. Basically end up back at the same message.

2 Likes
5

This is still occurring with 0.51.1 BTW

2 Likes
6

With the OpenRouter key, it still show the errors like:
The model anthropic/claude-3.5-sonnet does not work with your current plan or api key

But gpt-4o works well, and I try to downgrade the old version of cursor, and it shows the same error, probably the error is triggered by the Cursor API.

1 Like
7

hi @deanrie , do we have news about this issue?

thanks!

1 Like
8

got the same issue. Any progress? still occurring with Version: 1.0.0

2 Likes
9

I think they might be doing this on purpose, to force you use the pro version and not your own API keys. For the moment I moved to Cline which allows you to use your OpenRouter API keys without any problem.

10

i think you are right, they force user to pay to use their own service

1 Like
11

I’m now paying for the pro version and the custom API keys with OpenRouter still don’t work… I don’t understand why do they offer this option if it doesn’t work…

1 Like
12

Any news on this? I’m still having same issue

1 Like
13

i am experiencing this issue as well. using azure openai key and cursor 1.1.3.

14

image
image263×152 5.53 KB

image
image479×126 5.37 KB

below steps were my workarounds
turn off auto
selecting the model etc worked for me

15

Me too. I’ve tried my deepseek api but recieved a “Invalid Model”.

16

bump this.. it realy anoying.. i now going back to vscode and setup open router api in there.. it is sad.. i really liked cursor.. but seems they ignore this issue

17

still not fixed..

18

This topic was automatically closed 30 days after the last reply. New replies are no longer allowed.