feat: add supportsTemperature and Responses API flags

Extend ModelInfo schema with supportsTemperature and usesResponsesApi
capabilities to control request param inclusion and API selection.

Refactor OpenAiNativeHandler to generically handle Responses API models
instead of hardcoded families, normalizing IDs and gating temperature,
verbosity, and max token params via getModelParams.

Update GlamaHandler, LiteLLMHandler, UnboundHandler, and XAIHandler to
use getModelParams for capability-aware temperature/max token handling.

Enhance tests to cover Responses API flows, conversation continuity,
and temperature stripping for unsupported models, replacing SSE mocks
with SDK responses.create where applicable.

Author: snipeship

packages/types/src/model.ts

@@ -46,6 +46,10 @@ export const modelInfoSchema = z.object({
 	supportsPromptCache: z.boolean(),
 	// Capability flag to indicate whether the model supports an output verbosity parameter
 	supportsVerbosity: z.boolean().optional(),
+	// Indicates whether the model accepts a temperature parameter
+	supportsTemperature: z.boolean().optional(),
+	// Indicates that this model should be called via the Responses API instead of Chat Completions
+	usesResponsesApi: z.boolean().optional(),
 	supportsReasoningBudget: z.boolean().optional(),
 	requiredReasoningBudget: z.boolean().optional(),
 	supportsReasoningEffort: z.boolean().optional(),

packages/types/src/providers/openai.ts

@@ -19,6 +19,10 @@ 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,
+		usesResponsesApi: true,
+		// Q: Why do we not send the temperature for GPT-5?
+		// A: Because OpenAI does not support temperature over the API for GPT-5.
+		supportsTemperature: false,
 	},
 	"gpt-5-mini-2025-08-07": {
 		maxTokens: 128000,
@@ -32,6 +36,8 @@ 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,
+		usesResponsesApi: true,
+		supportsTemperature: false,
 	},
 	"gpt-5-nano-2025-08-07": {
 		maxTokens: 128000,
@@ -45,6 +51,8 @@ export const openAiNativeModels = {
 		cacheReadsPrice: 0.01,
 		description: "GPT-5 Nano: Fastest, most cost-efficient version of GPT-5",
 		supportsVerbosity: true,
+		usesResponsesApi: true,
+		supportsTemperature: false,
 	},
 	"gpt-4.1": {
 		maxTokens: 32_768,
@@ -83,6 +91,8 @@ export const openAiNativeModels = {
 		cacheReadsPrice: 0.5,
 		supportsReasoningEffort: true,
 		reasoningEffort: "medium",
+		usesResponsesApi: true,
+		supportsTemperature: false,
 	},
 	"o3-high": {
 		maxTokens: 100_000,
@@ -93,6 +103,8 @@ export const openAiNativeModels = {
 		outputPrice: 8.0,
 		cacheReadsPrice: 0.5,
 		reasoningEffort: "high",
+		usesResponsesApi: true,
+		supportsTemperature: false,
 	},
 	"o3-low": {
 		maxTokens: 100_000,
@@ -103,6 +115,8 @@ export const openAiNativeModels = {
 		outputPrice: 8.0,
 		cacheReadsPrice: 0.5,
 		reasoningEffort: "low",
+		usesResponsesApi: true,
+		supportsTemperature: false,
 	},
 	"o4-mini": {
 		maxTokens: 100_000,
@@ -114,6 +128,8 @@ export const openAiNativeModels = {
 		cacheReadsPrice: 0.275,
 		supportsReasoningEffort: true,
 		reasoningEffort: "medium",
+		usesResponsesApi: true,
+		supportsTemperature: false,
 	},
 	"o4-mini-high": {
 		maxTokens: 100_000,
@@ -124,6 +140,8 @@ export const openAiNativeModels = {
 		outputPrice: 4.4,
 		cacheReadsPrice: 0.275,
 		reasoningEffort: "high",
+		usesResponsesApi: true,
+		supportsTemperature: false,
 	},
 	"o4-mini-low": {
 		maxTokens: 100_000,
@@ -134,6 +152,8 @@ export const openAiNativeModels = {
 		outputPrice: 4.4,
 		cacheReadsPrice: 0.275,
 		reasoningEffort: "low",
+		usesResponsesApi: true,
+		supportsTemperature: false,
 	},
 	"o3-mini": {
 		maxTokens: 100_000,
@@ -145,6 +165,8 @@ export const openAiNativeModels = {
 		cacheReadsPrice: 0.55,
 		supportsReasoningEffort: true,
 		reasoningEffort: "medium",
+		usesResponsesApi: true,
+		supportsTemperature: false,
 	},
 	"o3-mini-high": {
 		maxTokens: 100_000,
@@ -155,6 +177,8 @@ export const openAiNativeModels = {
 		outputPrice: 4.4,
 		cacheReadsPrice: 0.55,
 		reasoningEffort: "high",
+		usesResponsesApi: true,
+		supportsTemperature: false,
 	},
 	"o3-mini-low": {
 		maxTokens: 100_000,
@@ -165,6 +189,8 @@ export const openAiNativeModels = {
 		outputPrice: 4.4,
 		cacheReadsPrice: 0.55,
 		reasoningEffort: "low",
+		usesResponsesApi: true,
+		supportsTemperature: false,
 	},
 	o1: {
 		maxTokens: 100_000,
@@ -174,6 +200,8 @@ export const openAiNativeModels = {
 		inputPrice: 15,
 		outputPrice: 60,
 		cacheReadsPrice: 7.5,
+		usesResponsesApi: true,
+		supportsTemperature: false,
 	},
 	"o1-preview": {
 		maxTokens: 32_768,
@@ -183,6 +211,8 @@ export const openAiNativeModels = {
 		inputPrice: 15,
 		outputPrice: 60,
 		cacheReadsPrice: 7.5,
+		usesResponsesApi: true,
+		supportsTemperature: false,
 	},
 	"o1-mini": {
 		maxTokens: 65_536,
@@ -192,6 +222,8 @@ export const openAiNativeModels = {
 		inputPrice: 1.1,
 		outputPrice: 4.4,
 		cacheReadsPrice: 0.55,
+		usesResponsesApi: true,
+		supportsTemperature: false,
 	},
 	"gpt-4.5-preview": {
 		maxTokens: 16_384,
@@ -228,6 +260,7 @@ export const openAiNativeModels = {
 		inputPrice: 1.5,
 		outputPrice: 6,
 		cacheReadsPrice: 0,
+		usesResponsesApi: true,
 		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.",
 	},
@@ -247,6 +280,5 @@ export const openAiModelInfoSaneDefaults: ModelInfo = {
 export const azureOpenAiDefaultApiVersion = "2024-08-01-preview"
 
 export const OPENAI_NATIVE_DEFAULT_TEMPERATURE = 0
-export const GPT5_DEFAULT_TEMPERATURE = 1.0
 
 export const OPENAI_AZURE_AI_INFERENCE_PATH = "/models/chat/completions"