this shows when i try to prompt anything using the kimi k2., “We’re having trouble connecting to the model provider. This might be temporary - please try again in a moment.”
kimi k2.5 and its a shame that cursor dont offer k2.6
For AI issues: add Request ID with privacy disabled
Request ID: 8a239bff-7198-4298-b1c2-d87b1da84111
{“error”:“ERROR_PROVIDER_ERROR”,“details”:{“title”:“Provider Error”,“detail”:“We’re having trouble connecting to the model provider. This might be temporary - please try again in a moment.”,“isRetryable”:false,“additionalInfo”:{},“buttons”:,“planChoices”:},“isExpected”:true}
Provider Error We’re having trouble connecting to the model provider. This might be temporary - please try again in a moment.
IAi: Provider Error We’re having trouble connecting to the model provider. This might be temporary - please try again in a moment.
at c5A (vscode-file://vscode-app/tmp/.mount_CursorYEutKv/usr/share/cursor/resources/app/out/vs/workbench/workbench.desktop.main.js:34247:23670)
at o5A (vscode-file://vscode-app/tmp/.mount_CursorYEutKv/usr/share/cursor/resources/app/out/vs/workbench/workbench.desktop.main.js:34247:22658)
at g5A (vscode-file://vscode-app/tmp/.mount_CursorYEutKv/usr/share/cursor/resources/app/out/vs/workbench/workbench.desktop.main.js:34248:6285)
at xou.run (vscode-file://vscode-app/tmp/.mount_CursorYEutKv/usr/share/cursor/resources/app/out/vs/workbench/workbench.desktop.main.js:34248:10400)
at async COa.runAgentLoop (vscode-file://vscode-app/tmp/.mount_CursorYEutKv/usr/share/cursor/resources/app/out/vs/workbench/workbench.desktop.main.js:46811:10024)
at async F3u.streamFromAgentBackend (vscode-file://vscode-app/tmp/.mount_CursorYEutKv/usr/share/cursor/resources/app/out/vs/workbench/workbench.desktop.main.js:46865:9277)
at async F3u.getAgentStreamResponse (vscode-file://vscode-app/tmp/.mount_CursorYEutKv/usr/share/cursor/resources/app/out/vs/workbench/workbench.desktop.main.js:46865:13663)
at async bMe.submitChatMaybeAbortCurrent (vscode-file://vscode-app/tmp/.mount_CursorYEutKv/usr/share/cursor/resources/app/out/vs/workbench/workbench.desktop.main.js:34308:17597)
at async Object.Ea [as onSubmit] (vscode-file://vscode-app/tmp/.mount_CursorYEutKv/usr/share/cursor/resources/app/out/vs/workbench/workbench.desktop.main.js:45790:4826)
at async vscode-file://vscode-app/tmp/.mount_CursorYEutKv/usr/share/cursor/resources/app/out/vs/workbench/workbench.desktop.main.js:45764:56755
Hey, thanks for the report and the Request ID. That made it easy to trace the request.
Two things:
There is an elevated error rate on kimi-k2.5 today on the provider side Fireworks. The team is aware and monitoring it. I don’t have an ETA for stabilization yet.
Your request failed for a different reason. The chat has 258+ messages, and one of them was sent in a format Kimi via Fireworks doesn’t accept. It was a multipart array instead of a plain string. The most reliable workaround is to start a new chat. Long threads on Kimi are generally unstable right now, so it’s better to keep the history shorter or switch to a different model for now.
For K2.6, there is a separate feature request thread where you can add a vote or comment: Kimi K2.6 Released - Request
Tried in new chat and its still not working also tired it after and before updating to the latest version and its still not working while works perfect fine on other alternatives of cursor and the funny thing is K2.6 isnt also available so cursor might be deliberately making sure people dont use the cheap models to get benefit, sadly.
Request ID: 9b5a6c08-ff16-40d5-878a-41ddc78032d5
{“error”:“ERROR_PROVIDER_ERROR”,“details”:{“title”:“Provider Error”,“detail”:“We’re having trouble connecting to the model provider. This might be temporary - please try again in a moment.”,“isRetryable”:false,“additionalInfo”:{},“buttons”:,“planChoices”:},“isExpected”:true}
Provider Error We’re having trouble connecting to the model provider. This might be temporary - please try again in a moment.
y7i: Provider Error We’re having trouble connecting to the model provider. This might be temporary - please try again in a moment.
at Fpk (vscode-file://vscode-app/usr/share/cursor/resources/app/out/vs/workbench/workbench.desktop.main.js:35852:26696)
at NDw (vscode-file://vscode-app/usr/share/cursor/resources/app/out/vs/workbench/workbench.desktop.main.js:35852:25580)
at Gpk (vscode-file://vscode-app/usr/share/cursor/resources/app/out/vs/workbench/workbench.desktop.main.js:35853:6852)
at hmf.run (vscode-file://vscode-app/usr/share/cursor/resources/app/out/vs/workbench/workbench.desktop.main.js:35853:11731)
at async qZs.runAgentLoop (vscode-file://vscode-app/usr/share/cursor/resources/app/out/vs/workbench/workbench.desktop.main.js:48220:22008)
at async OKf.streamFromAgentBackend (vscode-file://vscode-app/usr/share/cursor/resources/app/out/vs/workbench/workbench.desktop.main.js:48288:12345)
at async OKf.getAgentStreamResponse (vscode-file://vscode-app/usr/share/cursor/resources/app/out/vs/workbench/workbench.desktop.main.js:48288:19576)
at async Vyn.submitChatMaybeAbortCurrent (vscode-file://vscode-app/usr/share/cursor/resources/app/out/vs/workbench/workbench.desktop.main.js:35985:17397)
at async Object.zh [as onSubmit] (vscode-file://vscode-app/usr/share/cursor/resources/app/out/vs/workbench/workbench.desktop.main.js:47271:3889)
at async vscode-file://vscode-app/usr/share/cursor/resources/app/out/vs/workbench/workbench.desktop.main.js:47270:102122
If it’s failing even in a fresh chat, then it’s not about chat length. That was the case in the previous message. It’s more likely provider instability. We run Kimi K2.5 through a third-party provider, Fireworks, and their error rate has been spiking lately, which is why you’re seeing Provider Error. This is intermittent on their side, and we don’t have an ETA for when it’ll be stable again.
About “we intentionally block cheap models”, that’s not true. There’s no intent here. The model is just unstable on the provider side, which you can also see from the fact that other models work. It’s not in our interest for it to fail either.
For now, the best option is to switch to another model when Kimi errors out, then try again later. There’s a vote and discussion for K2.6 here: Kimi K2.6 Released - Request
Let me know if it fails every time or only sometimes. That’ll help us understand the scope.
Why not change the provider instead because thats the easiest thing, because how come it works perfectly on other IDE’s but only makes issues with CURSOR sasdly.
Hey, about switching providers, it’s not as simple as it sounds. Model availability, bandwidth limits, and terms depend on who’s hosting it, so we can’t just switch providers on the fly.
About “it works in other IDEs”, other clients might be reaching Kimi through a different provider or sending the request in a different format, so the behavior can differ. For us, K2.5 goes through Fireworks, and that’s where the error rate sometimes spikes.
To figure out what exactly is failing for you right now on June 10, please send a fresh Request ID with Privacy Mode turned off from your latest failed attempt. The older IDs you sent were from May. With a new one, we can check whether it’s the same Provider Error or something new.
Also, does it fail every time or only sometimes? That’ll help us understand how widespread it is.
