I'm getting this error generating with every model, i don't use free models, and have almost $100 in credits ATM.
Please investigate urgently @surreal bison
#URGENT Rate limit exceeded: free-models-per-minute
wary matrix
wary matrix
Received a user report 40 minutes ago, I'm not 100% sure exactly when it started.
It's still happening
surreal bison
hm, do you fall back to any free models in your request? can you paste the request params here?
we haven't changed anything here today
wary matrix
I think the default fallback is configured to a free model, on OR settings. But could every single model be failing?
I'll change the fallback now
surreal bison
oh interesting, that could be it, possible that we haven't tested that to see if it triggers rate limiting
hm, it shouldn't though unless you're not specifying model or models and you're not passing in "route": "fallback"
wary matrix
I'm using langchainjs, i don't pass any OR specific parameters. I've changed the default fallback and its generating again now.
surreal bison
do you pass in model or models in the request body?
surreal bison
ah i suspect this might be due to us validating the JSON you're passing in for your request body (something we recently added, cc @visual lance ) and seeing that it's not valid
wary matrix
It started working again after changing the fallback model though, and using the model I send in modelName, not the fallback model
surreal bison
hm
wary matrix
Seems like somehow after it triggered the fallback's model rate-limit, it started failing every subsequent request with that fallback model configured.