Add GPT-4o as possible model (#11764)

Resolves: #11766

Release Notes:

- Add GPT-4o support (see: https://openai.com/index/hello-gpt-4o/).
GPT-4o is better and faster than 4-turbo, at half the price.

crates/assistant/src/assistant_panel.rs

@@ -771,7 +771,8 @@ impl AssistantPanel {
             LanguageModel::OpenAi(model) => LanguageModel::OpenAi(match &model {
                 open_ai::Model::ThreePointFiveTurbo => open_ai::Model::Four,
                 open_ai::Model::Four => open_ai::Model::FourTurbo,
-                open_ai::Model::FourTurbo => open_ai::Model::ThreePointFiveTurbo,
+                open_ai::Model::FourTurbo => open_ai::Model::FourOmni,
+                open_ai::Model::FourOmni => open_ai::Model::ThreePointFiveTurbo,
             }),
             LanguageModel::ZedDotDev(model) => LanguageModel::ZedDotDev(match &model {
                 ZedDotDevModel::Gpt3Point5Turbo => ZedDotDevModel::Gpt4,

crates/assistant/src/assistant_settings.rs

@@ -414,7 +414,7 @@ mod tests {
         assert_eq!(
             AssistantSettings::get_global(cx).provider,
             AssistantProvider::OpenAi {
-                default_model: OpenAiModel::FourTurbo,
+                default_model: OpenAiModel::FourOmni,
                 api_url: open_ai_url(),
                 low_speed_timeout_in_seconds: None,
             }
@@ -436,7 +436,7 @@ mod tests {
         assert_eq!(
             AssistantSettings::get_global(cx).provider,
             AssistantProvider::OpenAi {
-                default_model: OpenAiModel::FourTurbo,
+                default_model: OpenAiModel::FourOmni,
                 api_url: "test-url".into(),
                 low_speed_timeout_in_seconds: None,
             }