Skip to content

Fix GPT-5 Responses API issues with condensing and image support #7067

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 13 commits into from
Aug 28, 2025
Merged

Fix GPT-5 Responses API issues with condensing and image support #7067

Show file tree
Hide file tree
Changes from 6 commits
Commits
Show all changes
13 commits
Select commit Hold shift + click to select a range
aa68b76
fix: migrate all OpenAI models to Responses API with proper image sup…
daniel-lxs Aug 13, 2025
48fc8b4
Fix GPT-5 Responses API issues with condensing and image support
daniel-lxs Aug 13, 2025
aa6ddbd
fix: remove debug logging statements from GPT-5 implementation
daniel-lxs Aug 14, 2025
516f7c3
refactor: improve code quality based on PR review
daniel-lxs Aug 14, 2025
f948a87
fix: address code review feedback for GPT-5 implementation
roomote Aug 18, 2025
c9e4ba3
test: log task disposal to satisfy Task.dispose test
daniel-lxs Aug 18, 2025
78c2173
Update src/core/task/Task.ts
daniel-lxs Aug 18, 2025
c2ec3d9
fix: use console.log in Task.dispose to satisfy unit test
daniel-lxs Aug 18, 2025
e1af388
feat(openai-native): always send instructions with Responses API; rem…
hannesrudolph Aug 27, 2025
c4c0b17
Implement code changes to enhance functionality and improve performance
hannesrudolph Aug 28, 2025
2022e28
openai-native: gate text.verbosity by supportsVerbosity, remove DEBUG…
hannesrudolph Aug 28, 2025
e0b4f1a
fix: implement completePrompt using Responses API non-streaming mode
daniel-lxs Aug 28, 2025
82aa867
test: update tests to match completePrompt implementation
daniel-lxs Aug 28, 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
2 changes: 2 additions & 0 deletions packages/types/src/model.ts
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -47,6 +47,8 @@ export const modelInfoSchema = z.object({
// Capability flag to indicate whether the model supports an output verbosity parameter
supportsVerbosity: z.boolean().optional(),
supportsReasoningBudget: z.boolean().optional(),
// Capability flag to indicate whether the model supports temperature parameter
supportsTemperature: z.boolean().optional(),
requiredReasoningBudget: z.boolean().optional(),
supportsReasoningEffort: z.boolean().optional(),
supportedParameters: z.array(modelParametersSchema).optional(),
Expand Down
21 changes: 21 additions & 0 deletions packages/types/src/providers/openai.ts
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -31,6 +31,7 @@ export const openAiNativeModels = {
description: "GPT-5: The best model for coding and agentic tasks across domains",
// supportsVerbosity is a new capability; ensure ModelInfo includes it
supportsVerbosity: true,
supportsTemperature: false,
},
"gpt-5-mini-2025-08-07": {
maxTokens: 128000,
Expand All @@ -44,6 +45,7 @@ export const openAiNativeModels = {
cacheReadsPrice: 0.03,
description: "GPT-5 Mini: A faster, more cost-efficient version of GPT-5 for well-defined tasks",
supportsVerbosity: true,
supportsTemperature: false,
},
"gpt-5-nano-2025-08-07": {
maxTokens: 128000,
Expand All @@ -57,6 +59,7 @@ export const openAiNativeModels = {
cacheReadsPrice: 0.01,
description: "GPT-5 Nano: Fastest, most cost-efficient version of GPT-5",
supportsVerbosity: true,
supportsTemperature: false,
},
"gpt-4.1": {
maxTokens: 32_768,
Expand All @@ -66,6 +69,7 @@ export const openAiNativeModels = {
inputPrice: 2,
outputPrice: 8,
cacheReadsPrice: 0.5,
supportsTemperature: true,
},
"gpt-4.1-mini": {
maxTokens: 32_768,
Expand All @@ -75,6 +79,7 @@ export const openAiNativeModels = {
inputPrice: 0.4,
outputPrice: 1.6,
cacheReadsPrice: 0.1,
supportsTemperature: true,
},
"gpt-4.1-nano": {
maxTokens: 32_768,
Expand All @@ -84,6 +89,7 @@ export const openAiNativeModels = {
inputPrice: 0.1,
outputPrice: 0.4,
cacheReadsPrice: 0.025,
supportsTemperature: true,
},
o3: {
maxTokens: 100_000,
Expand All @@ -95,6 +101,7 @@ export const openAiNativeModels = {
cacheReadsPrice: 0.5,
supportsReasoningEffort: true,
reasoningEffort: "medium",
supportsTemperature: false,
},
"o3-high": {
maxTokens: 100_000,
Expand All @@ -105,6 +112,7 @@ export const openAiNativeModels = {
outputPrice: 8.0,
cacheReadsPrice: 0.5,
reasoningEffort: "high",
supportsTemperature: false,
},
"o3-low": {
maxTokens: 100_000,
Expand All @@ -115,6 +123,7 @@ export const openAiNativeModels = {
outputPrice: 8.0,
cacheReadsPrice: 0.5,
reasoningEffort: "low",
supportsTemperature: false,
},
"o4-mini": {
maxTokens: 100_000,
Expand All @@ -126,6 +135,7 @@ export const openAiNativeModels = {
cacheReadsPrice: 0.275,
supportsReasoningEffort: true,
reasoningEffort: "medium",
supportsTemperature: false,
},
"o4-mini-high": {
maxTokens: 100_000,
Expand All @@ -136,6 +146,7 @@ export const openAiNativeModels = {
outputPrice: 4.4,
cacheReadsPrice: 0.275,
reasoningEffort: "high",
supportsTemperature: false,
},
"o4-mini-low": {
maxTokens: 100_000,
Expand All @@ -146,6 +157,7 @@ export const openAiNativeModels = {
outputPrice: 4.4,
cacheReadsPrice: 0.275,
reasoningEffort: "low",
supportsTemperature: false,
},
"o3-mini": {
maxTokens: 100_000,
Expand All @@ -157,6 +169,7 @@ export const openAiNativeModels = {
cacheReadsPrice: 0.55,
supportsReasoningEffort: true,
reasoningEffort: "medium",
supportsTemperature: false,
},
"o3-mini-high": {
maxTokens: 100_000,
Expand All @@ -167,6 +180,7 @@ export const openAiNativeModels = {
outputPrice: 4.4,
cacheReadsPrice: 0.55,
reasoningEffort: "high",
supportsTemperature: false,
},
"o3-mini-low": {
maxTokens: 100_000,
Expand All @@ -177,6 +191,7 @@ export const openAiNativeModels = {
outputPrice: 4.4,
cacheReadsPrice: 0.55,
reasoningEffort: "low",
supportsTemperature: false,
},
o1: {
maxTokens: 100_000,
Expand All @@ -186,6 +201,7 @@ export const openAiNativeModels = {
inputPrice: 15,
outputPrice: 60,
cacheReadsPrice: 7.5,
supportsTemperature: false,
},
"o1-preview": {
maxTokens: 32_768,
Expand All @@ -195,6 +211,7 @@ export const openAiNativeModels = {
inputPrice: 15,
outputPrice: 60,
cacheReadsPrice: 7.5,
supportsTemperature: false,
},
"o1-mini": {
maxTokens: 65_536,
Expand All @@ -204,6 +221,7 @@ export const openAiNativeModels = {
inputPrice: 1.1,
outputPrice: 4.4,
cacheReadsPrice: 0.55,
supportsTemperature: false,
},
"gpt-4o": {
maxTokens: 16_384,
Expand All @@ -213,6 +231,7 @@ export const openAiNativeModels = {
inputPrice: 2.5,
outputPrice: 10,
cacheReadsPrice: 1.25,
supportsTemperature: true,
},
"gpt-4o-mini": {
maxTokens: 16_384,
Expand All @@ -222,6 +241,7 @@ export const openAiNativeModels = {
inputPrice: 0.15,
outputPrice: 0.6,
cacheReadsPrice: 0.075,
supportsTemperature: true,
},
"codex-mini-latest": {
maxTokens: 16_384,
Expand All @@ -231,6 +251,7 @@ export const openAiNativeModels = {
inputPrice: 1.5,
outputPrice: 6,
cacheReadsPrice: 0,
supportsTemperature: false,
description:
"Codex Mini: Cloud-based software engineering agent powered by codex-1, a version of o3 optimized for coding tasks. Trained with reinforcement learning to generate human-style code, adhere to instructions, and iteratively run tests.",
},
Expand Down
10 changes: 5 additions & 5 deletions pnpm-lock.yaml
View file
Open in desktop

Some generated files are not rendered by default. Learn more about how customized files appear on GitHub.

8 changes: 8 additions & 0 deletions src/api/index.ts
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -52,6 +52,14 @@ export interface ApiHandlerCreateMessageMetadata {
* Used to enforce "skip once" after a condense operation.
*/
suppressPreviousResponseId?: boolean
/**
* Controls whether the response should be stored for 30 days in OpenAI's Responses API.
* When true (default), responses are stored and can be referenced in future requests
* using the previous_response_id for efficient conversation continuity.
* Set to false to opt out of response storage for privacy or compliance reasons.
* @default true
*/
store?: boolean
}

export interface ApiHandler {
Expand Down
Loading
Loading