feat: add conversation compaction support to Responses API

client-sdks/stainless/config.yml

@@ -175,6 +175,10 @@ resources:
     models:
       response_object_stream: OpenAIResponseObjectStream
       response_object: OpenAIResponseObject
+      compacted_response: OpenAICompactedResponse
+      response_input: OpenAIResponseInput
+      response_message: OpenAIResponseMessage
+      response_output: OpenAIResponseOutput
     methods:
       create:
         type: http
@@ -189,6 +193,9 @@ resources:
       delete:
         type: http
         endpoint: delete /v1/responses/{response_id}
+      compact:
+        type: http
+        endpoint: post /v1/responses/compact
     subresources:
       input_items:
         methods:

docs/docs/api-openai/conformance.mdx

@@ -18,20 +18,20 @@ This documentation is auto-generated from the OpenAI API specification compariso
 
 | Metric | Value |
 |--------|-------|
-| **Overall Conformance Score** | 87.8% |
-| **Endpoints Implemented** | 28/146 |
+| **Overall Conformance Score** | 87.6% |
+| **Endpoints Implemented** | 29/146 |
 | **Total Properties Checked** | 3441 |
-| **Schema/Type Issues** | 288 |
-| **Missing Properties** | 131 |
-| **Total Issues to Fix** | 419 |
+| **Schema/Type Issues** | 295 |
+| **Missing Properties** | 132 |
+| **Total Issues to Fix** | 427 |
 
 ## Integration Test Coverage
 
-**Overall Test Coverage Score: 44.1%**
+**Overall Test Coverage Score: 43.8%**
 
 | Category | Covered | Total | Score |
 |----------|---------|-------|-------|
-| CRUD Operations | 5 | 6 | 83.3% |
+| CRUD Operations | 5 | 7 | 71.4% |
 | Conversations | 5 | 9 | 55.6% |
 | Request Parameters | 21 | 25 | 84.0% |
 | Streaming Events | 16 | 53 | 30.2% |
@@ -51,7 +51,7 @@ Categories are sorted by conformance score (lowest first, needing most attention
 | Embeddings | 64.3% | 14 | 5 | 0 |
 | Files | 66.7% | 42 | 8 | 6 |
 | Models | 66.7% | 15 | 0 | 5 |
-| Responses | 86.7% | 225 | 29 | 1 |
+| Responses | 83.1% | 225 | 36 | 2 |
 | Chat | 87.1% | 403 | 33 | 19 |
 | Conversations | 98.8% | 2165 | 22 | 4 |
 
@@ -175,7 +175,6 @@ The following OpenAI API endpoints are not yet implemented in Llama Stack:
 
 ### /responses
 
-- `/responses/compact`
 - `/responses/input_tokens`
 
 ### /skills
@@ -964,7 +963,7 @@ Below is a detailed breakdown of conformance issues and missing properties for e
 
 ### Responses
 
-**Score:** 86.7% · **Issues:** 29 · **Missing:** 1
+**Score:** 83.1% · **Issues:** 36 · **Missing:** 2
 
 #### `/responses`
 
@@ -1014,6 +1013,32 @@ Below is a detailed breakdown of conformance issues and missing properties for e
 
 </details>
 
+#### `/responses/compact`
+
+**POST**
+
+<details>
+<summary>Missing Properties (1)</summary>
+
+- `requestBody.content.application/x-www-form-urlencoded`
+
+</details>
+
+<details>
+<summary>Schema Issues (7)</summary>
+
+| Property | Issues | Tested |
+|----------|--------|--------|
+| `requestBody.content.application/json.properties.input` | Union variants added: 2; Union variants removed: 1 | Yes |
+| `requestBody.content.application/json.properties.model` | Type added: ['string']; Union variants removed: 3 | Yes |
+| `responses.200.content.application/json.properties.object` | Default changed: response.compaction -> None | No |
+| `responses.200.content.application/json.properties.output.items` | Union variants added: 4 | Yes |
+| `responses.200.content.application/json.properties.usage` | Type removed: ['object'] | Yes |
+| `responses.200.content.application/json.properties.usage.properties.input_tokens_details` | Type removed: ['object'] | No |
+| `responses.200.content.application/json.properties.usage.properties.output_tokens_details` | Type removed: ['object'] | No |
+
+</details>
+
 ### Chat
 
 **Score:** 87.1% · **Issues:** 33 · **Missing:** 19

docs/docs/api-openai/provider_matrix.md

@@ -19,19 +19,19 @@ inference provider, based on integration test results.
 
 | Provider | Tested | Passing | Failing | Coverage |
 |----------|--------|---------|---------|----------|
-| azure | 102 | 102 | 0 | 86% |
-| bedrock | 25 | 25 | 0 | 21% |
-| openai | 119 | 119 | 0 | 100% |
+| azure | 113 | 113 | 0 | 87% |
+| bedrock | 25 | 25 | 0 | 19% |
+| openai | 130 | 130 | 0 | 100% |
 | vllm | 1 | 1 | 0 | 1% |
-| watsonx | 56 | 56 | 0 | 47% |
+| watsonx | 56 | 56 | 0 | 43% |
 
 ## Provider Details
 
 Models, endpoints, and versions used during test recordings.
 
 | Provider | Model(s) | Endpoint | Version Info |
 |----------|----------|----------|--------------|
-| azure | gpt-4o | llama-stack-test.openai.azure.com, lls-test.openai.azure.com | openai sdk: 2.5.0 |
+| azure | gpt-4o | llama-stack-test.openai.azure.com, lls-test.openai.azure.com | openai sdk: 2.30.0 |
 | bedrock | openai.gpt-oss-20b | bedrock-mantle.us-east-2.api.aws | openai sdk: 2.5.0 |
 | openai | gpt-4o, o4-mini, text-embedding-3-small | api.openai.com | openai sdk: 2.5.0 |
 | vllm | Qwen/Qwen3-0.6B | — | — |
@@ -53,6 +53,22 @@ Models, endpoints, and versions used during test recordings.
 | streaming basic | ✅ | ✅ | ✅ | — | ✅ |
 | streaming incremental content | ✅ | ✅ | ✅ | — | ✅ |
 
+## Compact Responses
+
+| Feature | azure | bedrock | openai | vllm | watsonx |
+| --- | --- | --- | --- | --- | --- |
+| compact basic conversation | ✅ | — | ✅ | — | — |
+| compact chain through compaction | ✅ | — | ✅ | — | — |
+| compact double compaction | ✅ | — | ✅ | — | — |
+| compact input items hides compaction | ✅ | — | ✅ | — | — |
+| compact roundtrip | ✅ | — | ✅ | — | — |
+| compact single message | ✅ | — | ✅ | — | — |
+| compact with previous response id | ✅ | — | ✅ | — | — |
+| compact with tool calls dropped | ✅ | — | ✅ | — | — |
+| context management auto compacts large input | ✅ | — | ✅ | — | — |
+| context management no compact below threshold | ✅ | — | ✅ | — | — |
+| context management none does not compact | ✅ | — | ✅ | — | — |
+
 ## Conversation Responses
 
 | Feature | azure | bedrock | openai | vllm | watsonx |