OpenAI o1-preview and o1-mini support (#17796)

Release Notes:

- Added support for OpenAI o1-mini and o1-preview models.

---------

Co-authored-by: Jason Mancuso <7891333+jvmncs@users.noreply.github.com>
Co-authored-by: Bennet <bennet@zed.dev>

crates/language_model/src/model/cloud_model.rs

@@ -102,6 +102,8 @@ impl CloudModel {
                 | open_ai::Model::FourTurbo
                 | open_ai::Model::FourOmni
                 | open_ai::Model::FourOmniMini
+                | open_ai::Model::O1Mini
+                | open_ai::Model::O1Preview
                 | open_ai::Model::Custom { .. } => {
                     LanguageModelAvailability::RequiresPlan(Plan::ZedPro)
                 }

crates/language_model/src/provider/open_ai.rs

@@ -372,10 +372,13 @@ pub fn count_open_ai_tokens(
                 })
                 .collect::<Vec<_>>();
 
-            if let open_ai::Model::Custom { .. } = model {
-                tiktoken_rs::num_tokens_from_messages("gpt-4", &messages)
-            } else {
-                tiktoken_rs::num_tokens_from_messages(model.id(), &messages)
+            match model {
+                open_ai::Model::Custom { .. }
+                | open_ai::Model::O1Mini
+                | open_ai::Model::O1Preview => {
+                    tiktoken_rs::num_tokens_from_messages("gpt-4", &messages)
+                }
+                _ => tiktoken_rs::num_tokens_from_messages(model.id(), &messages),
             }
         })
         .boxed()

crates/language_model/src/request.rs

@@ -241,6 +241,7 @@ pub struct LanguageModelRequest {
 
 impl LanguageModelRequest {
     pub fn into_open_ai(self, model: String, max_output_tokens: Option<u32>) -> open_ai::Request {
+        let stream = !model.starts_with("o1-");
         open_ai::Request {
             model,
             messages: self
@@ -259,7 +260,7 @@ impl LanguageModelRequest {
                     },
                 })
                 .collect(),
-            stream: true,
+            stream,
             stop: self.stop,
             temperature: self.temperature,
             max_tokens: max_output_tokens,