Responses API

POST https://api.shannon-ai.com/v1/responses

Ko te tono iti rawa he id model me tētahi kuputuhi. Kāore te whakautu e stream, mehemea kāore koe e tono.

from openai import OpenAI

client = OpenAI(
    api_key="YOUR_API_KEY",
    base_url="https://api.shannon-ai.com/v1",
)

response = client.responses.create(
    model="shannon-3",
    input="Say hello in one sentence.",
)

print(response.output_text)

import OpenAI from "openai";

const client = new OpenAI({
  apiKey: "YOUR_API_KEY",
  baseURL: "https://api.shannon-ai.com/v1",
});

const response = await client.responses.create({
  model: "shannon-3",
  input: "Say hello in one sentence.",
});

console.log(response.output_text);

curl https://api.shannon-ai.com/v1/responses \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "shannon-3",
    "input": "Say hello in one sentence."
  }'

He ahanoa whakautu te whakautu. Kei tōna output te reasoning, ina whakaaro te model, me te karere:

{
  "id": "resp_5f0c1e7a9b3d4c62a8e1f07d2b46c9a3",
  "object": "response",
  "created_at": 1791625200,
  "model": "shannon-3",
  "status": "completed",
  "error": null,
  "incomplete_details": null,
  "instructions": null,
  "metadata": {},
  "parallel_tool_calls": true,
  "temperature": null,
  "tool_choice": "auto",
  "tools": [],
  "top_p": null,
  "output": [
    {
      "id": "rs_8a2e6c1f4b7d4e03b9c5a1d70f3e2b64",
      "type": "reasoning",
      "summary": [
        {
          "type": "summary_text",
          "text": "The user wants a greeting in one sentence. Keep it short and friendly."
        }
      ]
    },
    {
      "id": "msg_1c7b9e3a5d2f4a86b0e4c8d19a6f7e25",
      "type": "message",
      "role": "assistant",
      "status": "completed",
      "content": [
        {
          "type": "output_text",
          "text": "Hello, it is good to meet you.",
          "annotations": []
        }
      ]
    }
  ],
  "usage": {
    "input_tokens": 1184,
    "output_tokens": 46,
    "total_tokens": 1230
  }
}

Ngā āpure tono

Ko te model me te input ngā mea e hiahiatia ana e tētahi tono. Ka whakaingoa te tīwae Applied by i ngā model e huri ai tētahi āpure i te whakautu. Ko ngā model open-weight hosted ngā id tekau mā rua o te rārangi model; ko te whānau Shannon 3 ko shannon-3, shannon-3-pro, shannon-3.1 me shannon-3.1-pro. Ngā model me ngā utu

E whakaaetia ana te previous_response_id, store, metadata, parallel_tool_calls, truncation, include, user me te prompt_cache_key kia rere tonu te waehere kiritaki o nāianei kāore i huri. Kāore rātou e huri i te whakautu. Nā te mea kāore he kōrero e tiakina ana, ka kawe te tono whai muri i ngā hurihanga o mua i te input.

Ngā tūemi input

He karere tētahi tūemi input, me tētahi role me tētahi content, he tūemi momo rānei e kawe ana i tētahi karanga tool, i tana hua rānei. He aho, he array wāhanga rānei te content.

Whakanōhia ngā tohutohu ki te āpure instructions. Ki tētahi id o te whānau Shannon 3, whakanōhia ngā tohutohu me puritia ki te karere user.

Ka taea te hoki o ngā tūemi o te output o tētahi whakautu ki te input pēnei i a rātou: ka pānuitia ngā tūemi message me function_call, ka whakarerea ngā tūemi reasoning.

Ngā wāhanga content

I runga i ngā taumata Shannon ka pānuitia ngā whakaahua me ngā kōnae mai i te tūemi user whakamutunga o te input. Kei tōna ake whārangi ngā rahi me ngā tepe. Ngā whakaahua me ngā kōnae

He tono me ngā tohutohu me ngā kōwhiringa

Ka tuku tēnei tono i ngā tohutohu, i tētahi kōrero hei tūemi input, i tētahi tepe output, i tētahi temperature me tētahi reasoning effort. Ka whakamahi i tētahi model open-weight hosted, ka whakamahi i ēnei katoa.

from openai import OpenAI

client = OpenAI(
    api_key="YOUR_API_KEY",
    base_url="https://api.shannon-ai.com/v1",
)

response = client.responses.create(
    model="DeepSeek-V4-Flash-0731-W4A16-AUTOROUND-REAP",
    instructions="You are a physics teacher. Answer in two sentences.",
    input=[
        {"role": "user", "content": "Why is the sky blue?"},
        {"role": "assistant", "content": "Air scatters blue light more than red light."},
        {"role": "user", "content": "And why are sunsets red?"},
    ],
    max_output_tokens=512,
    temperature=0.3,
    reasoning={"effort": "low"},
)

for item in response.output:
    if item.type == "reasoning":
        print("reasoning:", item.summary[0].text)
    elif item.type == "message":
        print("answer:", item.content[0].text)
print(response.usage)

import OpenAI from "openai";

const client = new OpenAI({
  apiKey: "YOUR_API_KEY",
  baseURL: "https://api.shannon-ai.com/v1",
});

const response = await client.responses.create({
  model: "DeepSeek-V4-Flash-0731-W4A16-AUTOROUND-REAP",
  instructions: "You are a physics teacher. Answer in two sentences.",
  input: [
    { role: "user", content: "Why is the sky blue?" },
    { role: "assistant", content: "Air scatters blue light more than red light." },
    { role: "user", content: "And why are sunsets red?" },
  ],
  max_output_tokens: 512,
  temperature: 0.3,
  reasoning: { effort: "low" },
});

for (const item of response.output) {
  if (item.type === "reasoning") console.log("reasoning:", item.summary[0].text);
  if (item.type === "message") console.log("answer:", item.content[0].text);
}
console.log(response.usage);

curl https://api.shannon-ai.com/v1/responses \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "DeepSeek-V4-Flash-0731-W4A16-AUTOROUND-REAP",
    "instructions": "You are a physics teacher. Answer in two sentences.",
    "input": [
      {"role": "user", "content": "Why is the sky blue?"},
      {"role": "assistant", "content": "Air scatters blue light more than red light."},
      {"role": "user", "content": "And why are sunsets red?"}
    ],
    "max_output_tokens": 512,
    "temperature": 0.3,
    "reasoning": {"effort": "low"}
  }'

Kei te whakautu a tētahi model open-weight hosted te id, object, model, created_at, status, output me te usage. Ka tāpiri tōna usage i ngā token i pānuitia mai i te cache me ngā token i whakapaua ki te reasoning:

{
  "usage": {
    "input_tokens": 58,
    "input_tokens_details": {
      "cached_tokens": 0
    },
    "output_tokens": 67,
    "output_tokens_details": {
      "reasoning_tokens": 19
    },
    "total_tokens": 125
  }
}

Te ahanoa whakautu

Kei ia whakautu te id, object, status, output me te usage. Hangaia tō waehere ki runga i ēnei.

Ngā tūemi output

Ka hono ngā SDK OpenAI i te kuputuhi o te karere mōu i te response.output_text.

Streaming

Tautuhia te stream ki te true mō ngā server-sent events. He rārangi event: tō ia frame me te ingoa o te kaupapa me tētahi rārangi data: me tētahi ahanoa JSON; ka tuaruatia e te ahanoa te ingoa i te type. Hono ngā delta ki tō rātou tūemi mā te output_index.

Ki te SDK OpenAI, pānuitia ngā kaupapa mā tō rātou type:

from openai import OpenAI

client = OpenAI(
    api_key="YOUR_API_KEY",
    base_url="https://api.shannon-ai.com/v1",
)

stream = client.responses.create(
    model="shannon-3",
    input="Explain server-sent events in three sentences.",
    stream=True,
)

for event in stream:
    kind = getattr(event, "type", "error")     # an event without a type is an error event
    if kind == "response.reasoning_summary_text.delta":
        pass                                   # a piece of reasoning: event.delta
    elif kind == "response.output_text.delta":
        print(event.delta, end="", flush=True)
    elif kind == "response.output_item.done" and event.item.type == "function_call":
        print("\ntool call:", event.item.name, event.item.arguments)
    elif kind == "response.completed":
        print("\n", event.response.usage)
    elif kind in ("response.failed", "error"):
        raise RuntimeError("the stream failed, send the request again")

import OpenAI from "openai";

const client = new OpenAI({
  apiKey: "YOUR_API_KEY",
  baseURL: "https://api.shannon-ai.com/v1",
});

const stream = await client.responses.create({
  model: "shannon-3",
  input: "Explain server-sent events in three sentences.",
  stream: true,
});

for await (const event of stream) {
  if (event.type === "response.reasoning_summary_text.delta") {
    // a piece of reasoning: event.delta
  } else if (event.type === "response.output_text.delta") {
    process.stdout.write(event.delta);
  } else if (event.type === "response.output_item.done" && event.item.type === "function_call") {
    console.log("\ntool call:", event.item.name, event.item.arguments);
  } else if (event.type === "response.completed") {
    console.log("\n", event.response.usage);
  } else if (event.type === "response.failed" || event.type === "error") {
    throw new Error("the stream failed, send the request again");
  }
}

curl -N https://api.shannon-ai.com/v1/responses \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "shannon-3",
    "input": "Explain server-sent events in three sentences.",
    "stream": true
  }'

Ka tae mai ngā kaupapa i tēnei raupapa. Ka puta ngā rōpū mō te reasoning, te karere me ngā karanga tool ina he ā rātou kei te whakautu anake.

event: response.created
data: {"type":"response.created","response":{"id":"resp_5f0c…","object":"response","status":"in_progress","output":[], …}}

event: response.in_progress
data: {"type":"response.in_progress","response":{"id":"resp_5f0c…","object":"response","status":"in_progress","output":[], …}}

event: response.output_item.added
data: {"type":"response.output_item.added","output_index":0,"item":{"id":"msg_1c7b…","type":"message","role":"assistant","status":"in_progress","content":[]}}

event: response.content_part.added
data: {"type":"response.content_part.added","output_index":0,"content_index":0,"part":{"type":"output_text","text":""}}

event: response.output_text.delta
data: {"type":"response.output_text.delta","output_index":0,"content_index":0,"delta":"Server-sent"}

event: response.output_text.delta
data: {"type":"response.output_text.delta","output_index":0,"content_index":0,"delta":" events let a server push text over one HTTP reply."}

event: response.output_text.done
data: {"type":"response.output_text.done","output_index":0,"content_index":0,"text":"Server-sent events let a server push text over one HTTP reply."}

event: response.content_part.done
data: {"type":"response.content_part.done","output_index":0,"content_index":0,"part":{"type":"output_text","text":"Server-sent events let a server push text over one HTTP reply."}}

event: response.output_item.done
data: {"type":"response.output_item.done","output_index":0,"item":{"id":"msg_1c7b…","type":"message","role":"assistant","status":"completed","content":[{"type":"output_text","text":"Server-sent events let a server push text over one HTTP reply.","annotations":[]}]}}

event: response.completed
data: {"type":"response.completed","response":{"id":"resp_5f0c…","object":"response","status":"completed","output":[…],"usage":{"input_tokens":1190,"output_tokens":18,"total_tokens":1208}, …}}

He stream ka rahua

I runga i ngā taumata Shannon ka mutu tētahi whakautu rahua me te response.failed hei whakakapi mō te response.completed. Ko tōna response.status he failed, ā, kei te response.error tētahi code, rate_limit_exceeded server_error rānei, me tētahi message.

I runga i ngā model open-weight hosted he kaupapa error te rahua me te code api_error me tētahi message. Ka whai tonu te response.completed i a ia.

Whakaarohia he rahua tētahi stream ina he response.failed he kaupapa error rānei kei roto, ā, tukuna anō te tono.

Ngā tool

He function te tool ka whakahaerehia e tō waehere. Whakaahuatia i te tools; ina hiahia te model ki a ia, kei te whakautu tētahi tūemi function_call hei whakakapi, i te taha rānei o te karere.

He whakamārama tool i te tools:

[
  {
    "type": "function",
    "name": "get_weather",
    "description": "Current weather for a city.",
    "parameters": {
      "type": "object",
      "properties": {
        "city": {
          "type": "string"
        }
      },
      "required": [
        "city"
      ]
    }
  }
]

Te output o tētahi whakautu e karanga ana i a ia:

[
  {
    "id": "call_3d9a7c1e5b2f4a60c8e1d7f09b24a6c5",
    "type": "function_call",
    "name": "get_weather",
    "call_id": "call_3d9a7c1e5b2f4a60c8e1d7f09b24a6c5",
    "arguments": "{\"city\":\"Paris\"}"
  }
]

Whakahaerehia te function, kātahi ka tuku anō i te tono me te karanga me tana hua kua tāpiritia ki te input:

[
  {
    "role": "user",
    "content": "What is the weather in Paris?"
  },
  {
    "type": "function_call",
    "call_id": "call_3d9a7c1e5b2f4a60c8e1d7f09b24a6c5",
    "name": "get_weather",
    "arguments": "{\"city\":\"Paris\"}"
  },
  {
    "type": "function_call_output",
    "call_id": "call_3d9a7c1e5b2f4a60c8e1d7f09b24a6c5",
    "output": "{\"temperature_c\": 18, \"sky\": \"cloudy\"}"
  }
]

Ka pānuitia ngā tool o te momo function anake. I runga i te shannon-1.6-*, shannon-2-* me te shannon-coder-1 ka whakaitia te schema parameters ki ngā kupu matua JSON Schema noa: ka pānuitia te oneOf hei anyOf, ā, kāore ngā kupu matua pēnei i te $ref, additionalProperties, const, default me te title e tukuna ki te model.

Ka whakamahia ngā uara tool_choice required me tētahi function whakaingoa e ngā model open-weight hosted.

Kua whakaahuatia katoatia te tool loop i tōna ake whārangi. Karanga ā‑mahi

Output whakaritea

Ka tuhi tēnei hōputu i te output whakaritea hei text.format. Kei te format tonu te schema, i te taha o te name me te strict, kaua i raro i tētahi mema json_schema pēnei i Chat Completions.

• Ka tono te {"type": "json_object"} i tētahi ahanoa JSON ahakoa tēhea.
• Ka tono te {"type": "json_schema", "name", "schema", "strict"} i tētahi JSON e whai ana i tō schema. Ko te strict he true ina waiho e koe.

{
  "model": "shannon-1.6-lite",
  "input": "Extract the city and the date: 'The conference opens in Lisbon on 12 May 2027.'",
  "text": {
    "format": {
      "type": "json_schema",
      "name": "event",
      "strict": true,
      "schema": {
        "type": "object",
        "properties": {
          "city": {
            "type": "string"
          },
          "date": {
            "type": "string",
            "description": "ISO 8601 date"
          }
        },
        "required": [
          "city",
          "date"
        ]
      }
    }
  }
}

Ka tae mai te JSON hei kuputuhi i te wāhanga output_text o te karere; parauhia mōu ake. Ka whakahoki tētahi model open-weight hosted kāore he output whakaritea i te 400.

{"city": "Lisbon", "date": "2027-05-12"}

Kua rārangihia i tōna ake whārangi ko tēhea model e whakaae ana ki tēhea āhua. Putanga hanganga

Reasoning

Ka whakahoki tētahi model e whakaaro ana i tana reasoning hei tūemi output o te momo reasoning, i mua i te karere. I roto i tētahi stream ka tae mai hei kaupapa response.reasoning_summary_text.delta.

Ka whakamahi ngā model open-weight hosted i te reasoning.effort: off, low, medium high rānei, ko high ina kāore koe e tuku i tētahi. Kāore te off e whakahoki i tētahi tūemi reasoning. Tukuna hei ahanoa; kāore e pānuitia tētahi aho kau i runga i tēnei endpoint.

Kua whakaahuatia ngā effort i tō rātou ake whārangi. Reasoning effort

Te tepe output

Ko te max_output_tokens he 4,096 ina kāore koe e tuku. Ko te maha o ngā token e wāhi atu ana i tō toenga ina tīmata te tono; ina oti te whakautu, ka whakakapia taua nui e te mea i whakamahia e te tono. Ka whakahoki tētahi uara kāore e uru ki tō toenga e toe ana i te 429 Quota exceeded.

Ka whakawhāiti i te roa o te whakautu i runga i te shannon-1.6-*, shannon-2-* me te shannon-coder-1, me te roa o te kuputuhi whakautu i runga i ngā model open-weight hosted.

I runga i tēnei endpoint ka utua te shannon-coder-1 ki ngā token mai i tō toenga, pērā i ia model kē atu.

Ka taea te tae o tētahi tinana tono ki te 32 MiB. Ngā tepe me te toenga

Ngā hapa

Ka whakamahi ngā hapa i te tinana hapa OpenAI: he ahanoa error me te type me te message; ka kawe ētahi hoki i te code me te param me te uara null. Ka rārangi te ripanga i ngā mea e whakahokia pinepine ana e tēnei endpoint. Kei tōna ake whārangi te rārangi katoa, me te mea hei whakamātau anō. Whakahaere hapa