shumengya/sproutclaw
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Align GPT-5.5 Codex capability handling (#3618)

Browse Source 
* fix(ai): align gpt-5.5 codex capabilities

* chore(ai): update generated models

* fix(ai): apply gpt-5.5 priority pricing to api

---------

Co-authored-by: Mario Zechner <badlogicgames@gmail.com>
This commit is contained in:
Markus Ylisiurunen
2026-04-24 00:49:33 +03:00
committed by GitHub
parent 86ba08a2e3
commit ba8f1c228a
5 changed files with 273 additions and 137 deletions
Download Patch File Download Diff File Expand all files Collapse all files

158
packages/ai/src/models.generated.ts
View File

@@ -2491,6 +2491,23 @@ export const MODELS = {
contextWindow: 1050000,
maxTokens: 128000,
} satisfies Model<"azure-openai-responses">,
"gpt-5.5": {
id: "gpt-5.5",
name: "GPT-5.5",
api: "azure-openai-responses",
provider: "azure-openai-responses",
baseUrl: "",
reasoning: true,
input: ["text", "image"],
cost: {
input: 5,
output: 30,
cacheRead: 0.5,
cacheWrite: 0,
},
contextWindow: 1050000,
maxTokens: 130000,
} satisfies Model<"azure-openai-responses">,
"o1": {
id: "o1",
name: "o1",
@@ -3894,8 +3911,8 @@ export const MODELS = {
contextWindow: 131072,
maxTokens: 8192,
} satisfies Model<"google-generative-ai">,
"gemma-4-26b-it": {
id: "gemma-4-26b-it",
"gemma-4-26b-a4b-it": {
id: "gemma-4-26b-a4b-it",
name: "Gemma 4 26B",
api: "google-generative-ai",
provider: "google",
@@ -6233,6 +6250,23 @@ export const MODELS = {
contextWindow: 1050000,
maxTokens: 128000,
} satisfies Model<"openai-responses">,
"gpt-5.5": {
id: "gpt-5.5",
name: "GPT-5.5",
api: "openai-responses",
provider: "openai",
baseUrl: "https://api.openai.com/v1",
reasoning: true,
input: ["text", "image"],
cost: {
input: 5,
output: 30,
cacheRead: 0.5,
cacheWrite: 0,
},
contextWindow: 1050000,
maxTokens: 130000,
} satisfies Model<"openai-responses">,
"o1": {
id: "o1",
name: "o1",
@@ -7019,6 +7053,23 @@ export const MODELS = {
contextWindow: 1050000,
maxTokens: 128000,
} satisfies Model<"openai-responses">,
"hy3-preview-free": {
id: "hy3-preview-free",
name: "Hy3 preview Free",
api: "openai-completions",
provider: "opencode",
baseUrl: "https://opencode.ai/zen/v1",
reasoning: true,
input: ["text"],
cost: {
input: 0,
output: 0,
cacheRead: 0,
cacheWrite: 0,
},
contextWindow: 256000,
maxTokens: 64000,
} satisfies Model<"openai-completions">,
"kimi-k2.5": {
id: "kimi-k2.5",
name: "Kimi K2.5",
@@ -7771,23 +7822,6 @@ export const MODELS = {
contextWindow: 131000,
maxTokens: 4096,
} satisfies Model<"openai-completions">,
"arcee-ai/trinity-large-preview:free": {
id: "arcee-ai/trinity-large-preview:free",
name: "Arcee AI: Trinity Large Preview (free)",
api: "openai-completions",
provider: "openrouter",
baseUrl: "https://openrouter.ai/api/v1",
reasoning: false,
input: ["text"],
cost: {
input: 0,
output: 0,
cacheRead: 0,
cacheWrite: 0,
},
contextWindow: 131000,
maxTokens: 4096,
} satisfies Model<"openai-completions">,
"arcee-ai/trinity-large-thinking": {
id: "arcee-ai/trinity-large-thinking",
name: "Arcee AI: Trinity Large Thinking",
@@ -8109,7 +8143,7 @@ export const MODELS = {
cacheWrite: 0,
},
contextWindow: 131072,
maxTokens: 32768,
maxTokens: 65536,
} satisfies Model<"openai-completions">,
"deepseek/deepseek-v3.2-exp": {
id: "deepseek/deepseek-v3.2-exp",
@@ -8159,7 +8193,7 @@ export const MODELS = {
cacheRead: 0.024999999999999998,
cacheWrite: 0.08333333333333334,
},
contextWindow: 1000000,
contextWindow: 1048576,
maxTokens: 8192,
} satisfies Model<"openai-completions">,
"google/gemini-2.0-flash-lite-001": {
@@ -8358,13 +8392,13 @@ export const MODELS = {
reasoning: true,
input: ["text", "image"],
cost: {
input: 0.07,
output: 0.35,
cacheRead: 0.04,
input: 0.06,
output: 0.33,
cacheRead: 0,
cacheWrite: 0,
},
contextWindow: 262144,
maxTokens: 262144,
maxTokens: 4096,
} satisfies Model<"openai-completions">,
"google/gemma-4-26b-a4b-it:free": {
id: "google/gemma-4-26b-a4b-it:free",
@@ -8394,7 +8428,7 @@ export const MODELS = {
cost: {
input: 0.13,
output: 0.38,
cacheRead: 0.019999999499999997,
cacheRead: 0,
cacheWrite: 0,
},
contextWindow: 262144,
@@ -8434,6 +8468,23 @@ export const MODELS = {
contextWindow: 128000,
maxTokens: 50000,
} satisfies Model<"openai-completions">,
"inclusionai/ling-2.6-1t:free": {
id: "inclusionai/ling-2.6-1t:free",
name: "inclusionAI: Ling-2.6-1T (free)",
api: "openai-completions",
provider: "openrouter",
baseUrl: "https://openrouter.ai/api/v1",
reasoning: false,
input: ["text"],
cost: {
input: 0,
output: 0,
cacheRead: 0,
cacheWrite: 0,
},
contextWindow: 262144,
maxTokens: 32768,
} satisfies Model<"openai-completions">,
"inclusionai/ling-2.6-flash:free": {
id: "inclusionai/ling-2.6-flash:free",
name: "inclusionAI: Ling-2.6-flash (free)",
@@ -8902,13 +8953,13 @@ export const MODELS = {
reasoning: false,
input: ["text"],
cost: {
input: 0.02,
output: 0.04,
input: 0.01,
output: 0.03,
cacheRead: 0,
cacheWrite: 0,
},
contextWindow: 131072,
maxTokens: 16384,
maxTokens: 4096,
} satisfies Model<"openai-completions">,
"mistralai/mistral-saba": {
id: "mistralai/mistral-saba",
@@ -9123,13 +9174,13 @@ export const MODELS = {
reasoning: true,
input: ["text", "image"],
cost: {
input: 0.7999999999999999,
output: 3.5,
cacheRead: 0.19999999999999998,
input: 0.7448,
output: 4.655,
cacheRead: 0.1463,
cacheWrite: 0,
},
contextWindow: 262144,
maxTokens: 262144,
contextWindow: 256000,
maxTokens: 65536,
} satisfies Model<"openai-completions">,
"nex-agi/deepseek-v3.1-nex-n1": {
id: "nex-agi/deepseek-v3.1-nex-n1",
@@ -10466,13 +10517,13 @@ export const MODELS = {
reasoning: true,
input: ["text"],
cost: {
input: 0.13,
output: 0.6,
input: 0.14950000000000002,
output: 1.495,
cacheRead: 0,
cacheWrite: 0,
},
contextWindow: 262144,
maxTokens: 262144,
contextWindow: 131072,
maxTokens: 4096,
} satisfies Model<"openai-completions">,
"qwen/qwen3-30b-a3b": {
id: "qwen/qwen3-30b-a3b",
@@ -11103,6 +11154,23 @@ export const MODELS = {
contextWindow: 262144,
maxTokens: 65536,
} satisfies Model<"openai-completions">,
"tencent/hy3-preview:free": {
id: "tencent/hy3-preview:free",
name: "Tencent: Hy3 preview (free)",
api: "openai-completions",
provider: "openrouter",
baseUrl: "https://openrouter.ai/api/v1",
reasoning: true,
input: ["text"],
cost: {
input: 0,
output: 0,
cacheRead: 0,
cacheWrite: 0,
},
contextWindow: 262144,
maxTokens: 262144,
} satisfies Model<"openai-completions">,
"thedrummer/rocinante-12b": {
id: "thedrummer/rocinante-12b",
name: "TheDrummer: Rocinante 12B",
@@ -11933,7 +12001,7 @@ export const MODELS = {
input: 0.5,
output: 3,
cacheRead: 0.09999999999999999,
cacheWrite: 0,
cacheWrite: 0.625,
},
contextWindow: 1000000,
maxTokens: 64000,
@@ -14055,9 +14123,9 @@ export const MODELS = {
reasoning: true,
input: ["text"],
cost: {
input: 0.09,
output: 0.29,
cacheRead: 0.045,
input: 0.09999999999999999,
output: 0.3,
cacheRead: 0.01,
cacheWrite: 0,
},
contextWindow: 262144,
@@ -14274,15 +14342,15 @@ export const MODELS = {
provider: "vercel-ai-gateway",
baseUrl: "https://ai-gateway.vercel.sh",
reasoning: true,
input: ["text", "image"],
input: ["text"],
cost: {
input: 1.4,
output: 4.4,
cacheRead: 0.26,
cacheWrite: 0,
},
contextWindow: 202752,
maxTokens: 202752,
contextWindow: 202800,
maxTokens: 64000,
} satisfies Model<"anthropic-messages">,
"zai/glm-5v-turbo": {
id: "zai/glm-5v-turbo",

24
packages/ai/src/providers/openai-codex-responses.ts
View File

@@ -356,26 +356,36 @@ function buildRequestBody(
function clampReasoningEffort(modelId: string, effort: string): string {
const id = modelId.includes("/") ? modelId.split("/").pop()! : modelId;
if ((id.startsWith("gpt-5.2") || id.startsWith("gpt-5.3") || id.startsWith("gpt-5.4")) && effort === "minimal")
if (
(id.startsWith("gpt-5.2") || id.startsWith("gpt-5.3") || id.startsWith("gpt-5.4") || id.startsWith("gpt-5.5")) &&
effort === "minimal"
)
return "low";
if (id === "gpt-5.1" && effort === "xhigh") return "high";
if (id === "gpt-5.1-codex-mini") return effort === "high" || effort === "xhigh" ? "high" : "medium";
return effort;
}
function getServiceTierCostMultiplier(serviceTier: ResponseCreateParamsStreaming["service_tier"] | undefined): number {
function getServiceTierCostMultiplier(
model: Pick<Model<"openai-codex-responses">, "id">,
serviceTier: ResponseCreateParamsStreaming["service_tier"] | undefined,
): number {
switch (serviceTier) {
case "flex":
return 0.5;
case "priority":
return 2;
return model.id === "gpt-5.5" ? 2.5 : 2;
default:
return 1;
}
}
function applyServiceTierPricing(usage: Usage, serviceTier: ResponseCreateParamsStreaming["service_tier"] | undefined) {
const multiplier = getServiceTierCostMultiplier(serviceTier);
function applyServiceTierPricing(
usage: Usage,
serviceTier: ResponseCreateParamsStreaming["service_tier"] | undefined,
model: Pick<Model<"openai-codex-responses">, "id">,
) {
const multiplier = getServiceTierCostMultiplier(model, serviceTier);
if (multiplier === 1) return;
usage.cost.input *= multiplier;
@@ -424,7 +434,7 @@ async function processStream(
await processResponsesStream(mapCodexEvents(parseSSE(response)), output, stream, model, {
serviceTier: options?.serviceTier,
resolveServiceTier: resolveCodexServiceTier,
applyServiceTierPricing,
applyServiceTierPricing: (usage, serviceTier) => applyServiceTierPricing(usage, serviceTier, model),
});
}
@@ -859,7 +869,7 @@ async function processWebSocketStream(
await processResponsesStream(mapCodexEvents(parseWebSocket(socket, options?.signal)), output, stream, model, {
serviceTier: options?.serviceTier,
resolveServiceTier: resolveCodexServiceTier,
applyServiceTierPricing,
applyServiceTierPricing: (usage, serviceTier) => applyServiceTierPricing(usage, serviceTier, model),
});
if (options?.signal?.aborted) {
keepConnection = false;

17
packages/ai/src/providers/openai-responses.ts
View File

@@ -106,7 +106,7 @@ export const streamOpenAIResponses: StreamFunction<"openai-responses", OpenAIRes
await processResponsesStream(openaiStream, output, stream, model, {
serviceTier: options?.serviceTier,
applyServiceTierPricing,
applyServiceTierPricing: (usage, serviceTier) => applyServiceTierPricing(usage, serviceTier, model),
});
if (options?.signal?.aborted) {
@@ -246,19 +246,26 @@ function buildParams(model: Model<"openai-responses">, context: Context, options
return params;
}
function getServiceTierCostMultiplier(serviceTier: ResponseCreateParamsStreaming["service_tier"] | undefined): number {
function getServiceTierCostMultiplier(
model: Pick<Model<"openai-responses">, "id">,
serviceTier: ResponseCreateParamsStreaming["service_tier"] | undefined,
): number {
switch (serviceTier) {
case "flex":
return 0.5;
case "priority":
return 2;
return model.id === "gpt-5.5" ? 2.5 : 2;
default:
return 1;
}
}
function applyServiceTierPricing(usage: Usage, serviceTier: ResponseCreateParamsStreaming["service_tier"] | undefined) {
const multiplier = getServiceTierCostMultiplier(serviceTier);
function applyServiceTierPricing(
usage: Usage,
serviceTier: ResponseCreateParamsStreaming["service_tier"] | undefined,
model: Pick<Model<"openai-responses">, "id">,
) {
const multiplier = getServiceTierCostMultiplier(model, serviceTier);
if (multiplier === 1) return;
usage.cost.input *= multiplier;

166
packages/ai/test/openai-codex-stream.test.ts
View File

@@ -406,6 +406,7 @@ describe("openai-codex streaming", () => {
await streamResult.result();
});
it.each(["gpt-5.3-codex", "gpt-5.4", "gpt-5.5"])("clamps %s minimal reasoning effort to low", async (modelId) => {
it("preserves gpt-5.5 xhigh reasoning effort from simple options", async () => {
const tempDir = mkdtempSync(join(tmpdir(), "pi-codex-stream-"));
process.env.PI_CODING_AGENT_DIR = tempDir;
@@ -558,93 +559,98 @@ describe("openai-codex streaming", () => {
});
it.each([
["flex", 0.5],
["priority", 2],
] as const)("uses the client-sent %s service tier when Codex echoes default", async (serviceTier, multiplier) => {
const tempDir = mkdtempSync(join(tmpdir(), "pi-codex-stream-"));
process.env.PI_CODING_AGENT_DIR = tempDir;
const token = mockToken();
const sse = `${[
`data: ${JSON.stringify({
type: "response.output_item.added",
item: { type: "message", id: "msg_1", role: "assistant", status: "in_progress", content: [] },
})}`,
`data: ${JSON.stringify({ type: "response.content_part.added", part: { type: "output_text", text: "" } })}`,
`data: ${JSON.stringify({ type: "response.output_text.delta", delta: "Hello" })}`,
`data: ${JSON.stringify({
type: "response.output_item.done",
item: {
type: "message",
id: "msg_1",
role: "assistant",
status: "completed",
content: [{ type: "output_text", text: "Hello" }],
},
})}`,
`data: ${JSON.stringify({
type: "response.completed",
response: {
status: "completed",
service_tier: "default",
usage: {
input_tokens: 1000000,
output_tokens: 1000000,
total_tokens: 2000000,
input_tokens_details: { cached_tokens: 0 },
["gpt-5.1-codex", "flex", 0.5],
["gpt-5.1-codex", "priority", 2],
["gpt-5.5", "flex", 0.5],
["gpt-5.5", "priority", 2.5],
] as const)(
"uses the client-sent %s service tier for %s when Codex echoes default",
async (modelId, serviceTier, multiplier) => {
const tempDir = mkdtempSync(join(tmpdir(), "pi-codex-stream-"));
process.env.PI_CODING_AGENT_DIR = tempDir;
const token = mockToken();
const sse = `${[
`data: ${JSON.stringify({
type: "response.output_item.added",
item: { type: "message", id: "msg_1", role: "assistant", status: "in_progress", content: [] },
})}`,
`data: ${JSON.stringify({ type: "response.content_part.added", part: { type: "output_text", text: "" } })}`,
`data: ${JSON.stringify({ type: "response.output_text.delta", delta: "Hello" })}`,
`data: ${JSON.stringify({
type: "response.output_item.done",
item: {
type: "message",
id: "msg_1",
role: "assistant",
status: "completed",
content: [{ type: "output_text", text: "Hello" }],
},
})}`,
`data: ${JSON.stringify({
type: "response.completed",
response: {
status: "completed",
service_tier: "default",
usage: {
input_tokens: 1000000,
output_tokens: 1000000,
total_tokens: 2000000,
input_tokens_details: { cached_tokens: 0 },
},
},
})}`,
].join("\n\n")}\n\n`;
const encoder = new TextEncoder();
const stream = new ReadableStream<Uint8Array>({
start(controller) {
controller.enqueue(encoder.encode(sse));
controller.close();
},
})}`,
].join("\n\n")}\n\n`;
});
const encoder = new TextEncoder();
const stream = new ReadableStream<Uint8Array>({
start(controller) {
controller.enqueue(encoder.encode(sse));
controller.close();
},
});
global.fetch = vi.fn(async (input: string | URL) => {
const url = typeof input === "string" ? input : input.toString();
if (url === "https://api.github.com/repos/openai/codex/releases/latest") {
return new Response(JSON.stringify({ tag_name: "rust-v0.0.0" }), { status: 200 });
}
if (url.startsWith("https://raw.githubusercontent.com/openai/codex/")) {
return new Response("PROMPT", { status: 200, headers: { etag: '"etag"' } });
}
if (url === "https://chatgpt.com/backend-api/codex/responses") {
return new Response(stream, {
status: 200,
headers: { "content-type": "text/event-stream" },
});
}
return new Response("not found", { status: 404 });
}) as typeof fetch;
global.fetch = vi.fn(async (input: string | URL) => {
const url = typeof input === "string" ? input : input.toString();
if (url === "https://api.github.com/repos/openai/codex/releases/latest") {
return new Response(JSON.stringify({ tag_name: "rust-v0.0.0" }), { status: 200 });
}
if (url.startsWith("https://raw.githubusercontent.com/openai/codex/")) {
return new Response("PROMPT", { status: 200, headers: { etag: '"etag"' } });
}
if (url === "https://chatgpt.com/backend-api/codex/responses") {
return new Response(stream, {
status: 200,
headers: { "content-type": "text/event-stream" },
});
}
return new Response("not found", { status: 404 });
}) as typeof fetch;
const model: Model<"openai-codex-responses"> = {
id: modelId,
name: modelId === "gpt-5.5" ? "GPT-5.5" : "GPT-5.1 Codex",
api: "openai-codex-responses",
provider: "openai-codex",
baseUrl: "https://chatgpt.com/backend-api",
reasoning: true,
input: ["text"],
cost: { input: 1, output: 2, cacheRead: 0, cacheWrite: 0 },
contextWindow: 400000,
maxTokens: 128000,
};
const model: Model<"openai-codex-responses"> = {
id: "gpt-5.1-codex",
name: "GPT-5.1 Codex",
api: "openai-codex-responses",
provider: "openai-codex",
baseUrl: "https://chatgpt.com/backend-api",
reasoning: true,
input: ["text"],
cost: { input: 1, output: 2, cacheRead: 0, cacheWrite: 0 },
contextWindow: 400000,
maxTokens: 128000,
};
const context: Context = {
systemPrompt: "You are a helpful assistant.",
messages: [{ role: "user", content: "Say hello", timestamp: Date.now() }],
};
const context: Context = {
systemPrompt: "You are a helpful assistant.",
messages: [{ role: "user", content: "Say hello", timestamp: Date.now() }],
};
const result = await streamOpenAICodexResponses(model, context, { apiKey: token, serviceTier }).result();
const result = await streamOpenAICodexResponses(model, context, { apiKey: token, serviceTier }).result();
expect(result.usage.cost.input).toBe(1 * multiplier);
expect(result.usage.cost.output).toBe(2 * multiplier);
expect(result.usage.cost.total).toBe(3 * multiplier);
});
expect(result.usage.cost.input).toBe(1 * multiplier);
expect(result.usage.cost.output).toBe(2 * multiplier);
expect(result.usage.cost.total).toBe(3 * multiplier);
},
);
it("does not set session_id/x-client-request-id headers when sessionId is not provided", async () => {
const tempDir = mkdtempSync(join(tmpdir(), "pi-codex-stream-"));

45
packages/ai/test/openai-responses-copilot-provider.test.ts
View File

@@ -137,4 +137,49 @@ describe("openai-responses provider defaults", () => {
expect(captured).toEqual({ sessionId: null, clientRequestId: null });
});
it.each([
["gpt-5.4", "priority", 2],
["gpt-5.5", "priority", 2.5],
["gpt-5.5", "flex", 0.5],
] as const)("applies %s %s service-tier cost multiplier", async (modelId, serviceTier, multiplier) => {
const model = getModel("openai", modelId);
const sse = `${[
`data: ${JSON.stringify({
type: "response.completed",
response: {
status: "completed",
service_tier: serviceTier,
usage: {
input_tokens: 1000000,
output_tokens: 1000000,
total_tokens: 2000000,
input_tokens_details: { cached_tokens: 0 },
},
},
})}`,
].join("\n\n")}\n\n`;
vi.spyOn(globalThis, "fetch").mockResolvedValue(
new Response(sse, {
status: 200,
headers: { "content-type": "text/event-stream" },
}),
);
const stream = streamOpenAIResponses(
model,
{
systemPrompt: "sys",
messages: [{ role: "user", content: "hi", timestamp: Date.now() }],
},
{ apiKey: "test-key", serviceTier },
);
const result = await stream.result();
expect(result.usage.cost.input).toBe(model.cost.input * multiplier);
expect(result.usage.cost.output).toBe(model.cost.output * multiplier);
expect(result.usage.cost.total).toBe((model.cost.input + model.cost.output) * multiplier);
});
});