Skip to content

Fallback to GPT-5.4-mini and cache the model used. #14369

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
sean-mcmanus merged 2 commits into from
Apr 13, 2026
Merged

Fallback to GPT-5.4-mini and cache the model used. #14369

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
f947dbe
Fallback to GPT-5.4-mini and cache the model used.
sean-mcmanus Apr 13, 2026
d03cd39
Fix ==.
sean-mcmanus Apr 13, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
31 changes: 24 additions & 7 deletions Extension/src/LanguageServer/Providers/CopilotHoverProvider.ts
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -27,6 +27,7 @@ export class CopilotHoverProvider implements vscode.HoverProvider {
private cancelledPosition: vscode.Position | undefined;
private content: string | undefined;
private chatModel: vscode.LanguageModelChat | undefined;
private chatModelId: string | undefined; // Save the selected model ID to avoid trying the same unavailable model repeatedly.
// Flag to avoid querying the LanguageModelChat repeatedly if no model is found
private checkedChatModel: boolean = false;
constructor(client: DefaultClient) {
Expand All @@ -41,13 +42,29 @@ export class CopilotHoverProvider implements vscode.HoverProvider {
const vscodelm = getVSCodeLanguageModel();
if (vscodelm) {
try {
// First look for GPT-4o which should be available to all
// users and have a 0x multiplier on paid plans.
// GTP-4o is faster than GPT-5-mini (which seems too slow for hover, e.g. 10+ seconds).
let [model] = await vscodelm.selectChatModels({ ...modelSelector, id: 'gpt-4o' });
if (!model) {
// If GPT-4o is not available, fall back to the first available model.
[model] = await vscodelm.selectChatModels(modelSelector);
let model: vscode.LanguageModelChat | undefined;
if (this.chatModelId === undefined) {
// First look for GPT-4o which should be available to all
// users and have a 0x multiplier on paid plans.
// GPT-4o is faster than the x0 GPT-5-mini (which seems too slow for hover, e.g. 10+ seconds).
this.chatModelId = 'gpt-4o';
[model] = await vscodelm.selectChatModels({ ...modelSelector, id: this.chatModelId });
if (!model) {
// If GPT-4o is not available, fallback to GPT-5.4-mini (x0.33 and fast).
this.chatModelId = 'gpt-5.4-mini';
[model] = await vscodelm.selectChatModels({ ...modelSelector, id: this.chatModelId });
}
if (!model) {
// If GPT-5.4-mini is not available, fallback to the first available model.
this.chatModelId = 'default';
[model] = await vscodelm.selectChatModels(modelSelector);
}
} else {
if (this.chatModelId === 'default') {
[model] = await vscodelm.selectChatModels(modelSelector);
} else {
[model] = await vscodelm.selectChatModels({ ...modelSelector, id: this.chatModelId });
}
}
if (!model) {
telemetry.logLanguageServerEvent('CopilotHoverNoModelSelected', { remoteName: vscode.env.remoteName || 'local' });
Expand Down
Loading