Skip to content

feat: add GPT-5 model support #6819

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
merged 17 commits into from
Aug 7, 2025
Merged

feat: add GPT-5 model support #6819

Show file tree
Hide file tree
Changes from 4 commits
Commits
Show all changes
17 commits
Select commit Hold shift + click to select a range
0e7abee
feat: add GPT-5 model support
roomote Aug 7, 2025
b99eb41
fix: remove hardcoded temperature from GPT-5 handler
roomote Aug 7, 2025
57eccb5
feat: add reasoning effort support for all OpenAI models
roomote Aug 7, 2025
4d6056e
fix: update test to expect new default model gpt-5-2025-08-07
roomote Aug 7, 2025
fe82301
feat: increase GPT-5 models context window to 400,000
roomote Aug 7, 2025
0c32602
revert: remove GPT-5 models, keep only nectarine experimental model
roomote Aug 7, 2025
7c5e5a0
feat: add GPT-5 models with updated context windows
roomote Aug 7, 2025
7251237
fix: restore reasoning effort support for o1 series models
hannesrudolph Aug 7, 2025
e37a6d2
Revert "fix: restore reasoning effort support for o1 series models"
hannesrudolph Aug 7, 2025
a75a2b8
fix: restore reasoning effort support for o3 and o4 models
hannesrudolph Aug 7, 2025
f229392
Revert "fix: restore reasoning effort support for o3 and o4 models"
hannesrudolph Aug 7, 2025
77b3be7
fix: restore reasoning effort support for o3 and o4 models
hannesrudolph Aug 7, 2025
4a3a7ac
fix: adjust reasoning effort support for o3/o4 models
hannesrudolph Aug 7, 2025
c402014
fix: remove nectarine experimental model
hannesrudolph Aug 7, 2025
3f9bd5d
feat: implement full GPT-5 support with verbosity and minimal reasoning
hannesrudolph Aug 7, 2025
abe3252
feat: add verbosity setting for GPT-5 models
daniel-lxs Aug 7, 2025
331ecf2
Delete .roorules
daniel-lxs Aug 7, 2025
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
11 changes: 11 additions & 0 deletions .changeset/gpt5-support.md
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,11 @@
---
"@roo-code/types": minor
"roo-cline": minor
---

Add GPT-5 model support

- Added GPT-5 models (gpt-5-2025-08-07, gpt-5-mini-2025-08-07, gpt-5-nano-2025-08-07) to OpenAI Native provider
- Added nectarine-alpha-new-reasoning-effort-2025-07-25 experimental model
- Set gpt-5-2025-08-07 as the new default OpenAI Native model
- Implemented GPT-5 specific handling with streaming and reasoning effort support
Copy link
Contributor Author

@roomote roomote bot Aug 7, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

The changeset mentions 'reasoning effort support' but I don't see the flag set for GPT-5 models like it is for o3/o4 models. Should we either add the flag or remove this from the description?

51 changes: 50 additions & 1 deletion packages/types/src/providers/openai.ts
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -3,14 +3,55 @@ import type { ModelInfo } from "../model.js"
// https://openai.com/api/pricing/
export type OpenAiNativeModelId = keyof typeof openAiNativeModels

export const openAiNativeDefaultModelId: OpenAiNativeModelId = "gpt-4.1"
export const openAiNativeDefaultModelId: OpenAiNativeModelId = "gpt-5-2025-08-07"

export const openAiNativeModels = {
"gpt-5-2025-08-07": {
maxTokens: 128000,
contextWindow: 256000,
supportsImages: true,
supportsPromptCache: true,
supportsReasoningEffort: true,
inputPrice: 1.25,
outputPrice: 10.0,
cacheReadsPrice: 0.125,
},
"gpt-5-mini-2025-08-07": {
maxTokens: 128000,
contextWindow: 256000,
supportsImages: true,
supportsPromptCache: true,
supportsReasoningEffort: true,
inputPrice: 0.25,
outputPrice: 2.0,
cacheReadsPrice: 0.025,
},
"gpt-5-nano-2025-08-07": {
maxTokens: 128000,
contextWindow: 256000,
supportsImages: true,
supportsPromptCache: true,
supportsReasoningEffort: true,
inputPrice: 0.05,
outputPrice: 0.4,
cacheReadsPrice: 0.005,
},
"nectarine-alpha-new-reasoning-effort-2025-07-25": {
maxTokens: 128000,
contextWindow: 256000,
supportsImages: true,
supportsPromptCache: true,
supportsReasoningEffort: true,
inputPrice: 0,
outputPrice: 0,
cacheReadsPrice: 0,
},
"gpt-4.1": {
maxTokens: 32_768,
contextWindow: 1_047_576,
supportsImages: true,
supportsPromptCache: true,
supportsReasoningEffort: true,
inputPrice: 2,
outputPrice: 8,
cacheReadsPrice: 0.5,
Expand All @@ -20,6 +61,7 @@ export const openAiNativeModels = {
contextWindow: 1_047_576,
supportsImages: true,
supportsPromptCache: true,
supportsReasoningEffort: true,
inputPrice: 0.4,
outputPrice: 1.6,
cacheReadsPrice: 0.1,
Expand All @@ -29,6 +71,7 @@ export const openAiNativeModels = {
contextWindow: 1_047_576,
supportsImages: true,
supportsPromptCache: true,
supportsReasoningEffort: true,
inputPrice: 0.1,
outputPrice: 0.4,
cacheReadsPrice: 0.025,
Expand Down Expand Up @@ -131,6 +174,7 @@ export const openAiNativeModels = {
contextWindow: 200_000,
supportsImages: true,
supportsPromptCache: true,
supportsReasoningEffort: true,
inputPrice: 15,
outputPrice: 60,
cacheReadsPrice: 7.5,
Expand All @@ -140,6 +184,7 @@ export const openAiNativeModels = {
contextWindow: 128_000,
supportsImages: true,
supportsPromptCache: true,
supportsReasoningEffort: true,
inputPrice: 15,
outputPrice: 60,
cacheReadsPrice: 7.5,
Expand All @@ -149,6 +194,7 @@ export const openAiNativeModels = {
contextWindow: 128_000,
supportsImages: true,
supportsPromptCache: true,
supportsReasoningEffort: true,
inputPrice: 1.1,
outputPrice: 4.4,
cacheReadsPrice: 0.55,
Expand All @@ -158,6 +204,7 @@ export const openAiNativeModels = {
contextWindow: 128_000,
supportsImages: true,
supportsPromptCache: true,
supportsReasoningEffort: true,
inputPrice: 75,
outputPrice: 150,
cacheReadsPrice: 37.5,
Expand All @@ -167,6 +214,7 @@ export const openAiNativeModels = {
contextWindow: 128_000,
supportsImages: true,
supportsPromptCache: true,
supportsReasoningEffort: true,
inputPrice: 2.5,
outputPrice: 10,
cacheReadsPrice: 1.25,
Expand All @@ -176,6 +224,7 @@ export const openAiNativeModels = {
contextWindow: 128_000,
supportsImages: true,
supportsPromptCache: true,
supportsReasoningEffort: true,
inputPrice: 0.15,
outputPrice: 0.6,
cacheReadsPrice: 0.075,
Expand Down
2 changes: 1 addition & 1 deletion src/api/providers/__tests__/openai-native.spec.ts
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -455,7 +455,7 @@ describe("OpenAiNativeHandler", () => {
openAiNativeApiKey: "test-api-key",
})
const modelInfo = handlerWithoutModel.getModel()
expect(modelInfo.id).toBe("gpt-4.1") // Default model
expect(modelInfo.id).toBe("gpt-5-2025-08-07") // Default model
expect(modelInfo.info).toBeDefined()
})
})
Expand Down
31 changes: 31 additions & 0 deletions src/api/providers/openai-native.ts
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -53,6 +53,8 @@ export class OpenAiNativeHandler extends BaseProvider implements SingleCompletio
yield* this.handleReasonerMessage(model, id, systemPrompt, messages)
} else if (model.id.startsWith("o1")) {
yield* this.handleO1FamilyMessage(model, systemPrompt, messages)
} else if (this.isGPT5Model(model.id)) {
yield* this.handleGPT5Message(model, systemPrompt, messages)
} else {
yield* this.handleDefaultModelMessage(model, systemPrompt, messages)
}
Expand All @@ -66,6 +68,8 @@ export class OpenAiNativeHandler extends BaseProvider implements SingleCompletio
// o1 supports developer prompt with formatting
// o1-preview and o1-mini only support user messages
const isOriginalO1 = model.id === "o1"
const { reasoning } = this.getModel()

const response = await this.client.chat.completions.create({
model: model.id,
messages: [
Expand All @@ -77,6 +81,7 @@ export class OpenAiNativeHandler extends BaseProvider implements SingleCompletio
],
stream: true,
stream_options: { include_usage: true },
...(reasoning && reasoning),
})

yield* this.handleStreamResponse(response, model)
Expand Down Expand Up @@ -112,17 +117,43 @@ export class OpenAiNativeHandler extends BaseProvider implements SingleCompletio
systemPrompt: string,
messages: Anthropic.Messages.MessageParam[],
): ApiStream {
const { reasoning } = this.getModel()

const stream = await this.client.chat.completions.create({
model: model.id,
temperature: this.options.modelTemperature ?? OPENAI_NATIVE_DEFAULT_TEMPERATURE,
messages: [{ role: "system", content: systemPrompt }, ...convertToOpenAiMessages(messages)],
stream: true,
stream_options: { include_usage: true },
...(reasoning && reasoning),
})

yield* this.handleStreamResponse(stream, model)
}

private async *handleGPT5Message(
model: OpenAiNativeModel,
systemPrompt: string,
messages: Anthropic.Messages.MessageParam[],
): ApiStream {
const { reasoning } = this.getModel()

const stream = await this.client.chat.completions.create({
model: model.id,
temperature: this.options.modelTemperature ?? OPENAI_NATIVE_DEFAULT_TEMPERATURE,
messages: [{ role: "developer", content: systemPrompt }, ...convertToOpenAiMessages(messages)],
stream: true,
stream_options: { include_usage: true },
...(reasoning && reasoning),
})

yield* this.handleStreamResponse(stream, model)
}

private isGPT5Model(modelId: string): boolean {
return modelId.includes("gpt-5") || modelId.includes("gpt5") || modelId.includes("nectarine")
Copy link
Contributor Author

@roomote roomote bot Aug 7, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This detection logic is quite broad - any model containing 'nectarine' would be treated as GPT-5. Could we be more explicit?

}

private async *handleStreamResponse(
stream: AsyncIterable<OpenAI.Chat.Completions.ChatCompletionChunk>,
model: OpenAiNativeModel,
Expand Down
Loading