[Feature] Add /v1/decisions for typed decisions on candidate scoring

docs/cookbook/autoregressive/Qwen/Qwen3.5.mdx

@@ -777,6 +777,37 @@ Tool Call: None
    Arguments: }
 ```
 
+#### 4.3.3 Semantic Decisions
+
+Qwen3.5-35B-A3B can answer typed questions about an input with a probability for every option through `/v1/decisions`, without generating text. The server turns thinking off for these requests, so the answer is read after a closed think block, and the same server keeps serving chat traffic. No launch flag is needed. Until a release contains `/v1/decisions`, install a [nightly build](/docs/get-started/install#nightly-builds). With a server started by `sglang serve --model-path Qwen/Qwen3.5-35B-A3B --host 127.0.0.1 --port 30000`:
+
+```python Example
+import requests
+
+response = requests.post(
+    "http://127.0.0.1:30000/v1/decisions",
+    json={
+        "input": "I've been trying to connect my Stripe account for 3 days and the integration keeps failing.",
+        "questions": [
+            {
+                "id": "team",
+                "type": "choice",
+                "question": "Which team should handle this ticket?",
+                "options": [{"name": "billing"}, {"name": "technical"}, {"name": "sales"}],
+            },
+            {"id": "urgent", "type": "yes_no", "question": "The customer needs an answer today."},
+        ],
+    },
+    timeout=60,
+)
+response.raise_for_status()
+answers = response.json()["answers"]
+print(answers["team"]["choice"], answers["team"]["probabilities"])
+print(answers["urgent"]["probabilities"]["yes"])
+```
+
+See [Decision models](/docs/supported-models/decision_models) for the request reference, how the probabilities are computed, and the limits.
+
 ## 5. Benchmark
 
 ### 5.1 Accuracy Benchmark

docs/cookbook/autoregressive/Qwen/Qwen3.8-27B.mdx

@@ -541,3 +541,34 @@ providers:
 ```
 
 </Accordion>
+
+## 4. Semantic Decisions
+
+Qwen3.8-27B can answer typed questions about an input with a probability for every option through `/v1/decisions`, without generating text. The server turns thinking off for these requests, so the answer is read after a closed think block, and the same server keeps serving chat traffic. No launch flag is needed. Until a release contains `/v1/decisions`, install a [nightly build](/docs/get-started/install#nightly-builds). With a server started by `sglang serve --model-path Qwen/Qwen3.8-27B --host 127.0.0.1 --port 30000`:
+
+```python Example
+import requests
+
+response = requests.post(
+    "http://127.0.0.1:30000/v1/decisions",
+    json={
+        "input": "I've been trying to connect my Stripe account for 3 days and the integration keeps failing.",
+        "questions": [
+            {
+                "id": "team",
+                "type": "choice",
+                "question": "Which team should handle this ticket?",
+                "options": [{"name": "billing"}, {"name": "technical"}, {"name": "sales"}],
+            },
+            {"id": "urgent", "type": "yes_no", "question": "The customer needs an answer today."},
+        ],
+    },
+    timeout=60,
+)
+response.raise_for_status()
+answers = response.json()["answers"]
+print(answers["team"]["choice"], answers["team"]["probabilities"])
+print(answers["urgent"]["probabilities"]["yes"])
+```
+
+See [Decision models](/docs/supported-models/decision_models) for the request reference, how the probabilities are computed, and the limits.

docs/docs.json

@@ -998,7 +998,8 @@
               {
                 "group": "Specialized Models",
                 "pages": [
-                  "docs/supported-models/reward_models"
+                  "docs/supported-models/reward_models",
+                  "docs/supported-models/decision_models"
                 ]
               },
               {

docs/docs/basic_usage/native_api.mdx

@@ -15,6 +15,7 @@ Apart from the OpenAI compatible APIs, the SGLang Runtime also provides its nati
 - `/encode`(embedding model)
 - `/v1/rerank`(cross encoder rerank model)
 - `/v1/score`(decoder-only scoring)
+- `/v1/decisions`(typed decisions on a chat model)
 - `/classify`(reward model)
 - `/start_expert_distribution_record`
 - `/stop_expert_distribution_record`
@@ -256,6 +257,8 @@ Parameters:
 - `item_first`: Whether items come first in concatenation order (default: False)
 - `model`: Model name
 
+`query` and `items` also accept token ids, with `query` set to `[]` when each item is a complete prompt. `label_token_ids` can be one list per item. `temperature` divides the label logits before the softmax and needs `apply_softmax`, and `return_token_logprobs` adds `token_logprobs`, the full-vocabulary log-probabilities of the labels.
+
 The response contains `scores` - a list of probability lists, one per item, each in the order of `label_token_ids`.
 
 ```python Example
@@ -293,6 +296,20 @@ for item, scores in zip(items, response_json["scores"]):
 terminate_process(score_process)
 ```
 
+## v1/decisions (typed decisions on a chat model)
+
+Answer typed questions about an input with per-option probabilities, without generating. The server renders each question with the model's chat template and thinking turned off, assigns one-token answer labels, and scores them through the same path as `v1/score`.
+
+Parameters:
+- `input`: Text, object, or array the questions are about
+- `questions`: List of questions, each with a unique `id` and a `type` of `choice` (with `options`), `score` (with `levels`), or `yes_no`
+- `temperature`: Divides the label logits before the softmax (default: 1)
+- `chat_template_kwargs`: Extra chat template arguments
+- `prompt_format_version`: Optional pin of the server-owned prompt wording
+- `return_prompt_token_ids`: Return the scored prompt and label token ids for replay through `v1/score` (default: False)
+
+The response contains `answers` keyed by question id, each with `type`, `probabilities`, and `label_mass`, plus `choice` for a choice question or `score` for a score question. A yes or no answer is read from `probabilities["yes"]`. See [Decision models](/docs/supported-models/decision_models) for examples, the full reference, and error cases.
+
 ## Classify (reward model)
 
 SGLang Runtime also supports reward models. Here we use a reward model to classify the quality of pairwise generations.

docs/docs/supported-models.mdx

@@ -84,4 +84,13 @@ SGLang supports model families across text generation, retrieval, and reward wor
   >
     RLHF and reward scoring pipelines optimized for production latency.
   </Card>
+  <Card
+    title="Decision models"
+    mode="card"
+    className="max-w-sm mx-auto"
+    href="./supported-models/decision_models"
+    icon="list-check"
+  >
+    Typed choice, score, and yes or no answers with a probability for every option.
+  </Card>
 </CardGroup>