Chat Completions

POST https://api.shannon-ai.com/v1/chat/completions

Is e an t-iarrtas as lugha id modail agus aon teachdaireachd neach-cleachdaidh.

from openai import OpenAI

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

response = client.chat.completions.create(
    model="shannon-3",
    messages=[{"role": "user", "content": "Say hello in one sentence."}],
)

print(response.choices[0].message.content)

import OpenAI from "openai";

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

const response = await client.chat.completions.create({
  model: "shannon-3",
  messages: [{ role: "user", content: "Say hello in one sentence." }],
});

console.log(response.choices[0].message.content);

curl https://api.shannon-ai.com/v1/chat/completions \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "shannon-3",
    "messages": [{"role": "user", "content": "Say hello in one sentence."}]
  }'

Is e aon nì JSON am freagairt:

{
  "id": "chatcmpl-5f0c1e7a9b3d4c62a8e1f07d2b46c9a3",
  "object": "chat.completion",
  "created": 1791625200,
  "model": "shannon-3",
  "choices": [
    {
      "index": 0,
      "message": {
        "role": "assistant",
        "content": "Hello, it is good to meet you.",
        "reasoning_content": "The user wants a greeting in one sentence. Keep it short and friendly."
      },
      "finish_reason": "stop"
    }
  ],
  "usage": {
    "prompt_tokens": 1184,
    "completion_tokens": 46,
    "total_tokens": 1230
  }
}

Headers

Headers an iarrtais

Headers na freagairt

Raointean an iarrtais

Chan eil ach messages riatanach. Ainmichidh a’ cholbh Air a chur an sàs le na modailean air an atharraich raon am freagairt. Is e na modailean open-weight hosted an dusan id air liosta nam modailean; is e teaghlach Shannon 3 shannon-3, shannon-3-pro, shannon-3.1 agus shannon-3.1-pro. Modailean & prìsean

Thèid raointean OpenAI eile, leithid n, user, stream_options, parallel_tool_calls, presence_penalty, frequency_penalty, logit_bias, logprobs, metadata, store agus prompt_cache_key, aontachadh gus an ruith còd client a th’ ann mar a tha e. Chan atharraich iad am freagairt: tha aon roghainn ann an-còmhnaidh, agus bidh usage aig deireadh stream an-còmhnaidh.

Tillidh raon le seòrsa JSON ceàrr, mar eisimpleir "max_tokens": "100", 422. Mar sin cuideachd iarrtas gun messages.

Tha duilleag fhèin aig innealan, output structaraichte, reusanachadh agus rannsachadh lìn: Call gnìomh, Toraidhean structarail, Oidhirp reusanachaidh, Rannsachadh Lìn air a thoirt a-steach.

Iarrtas le roghainnean

Suidhichidh an t-iarrtas seo teachdaireachd system, na raointean samplachaidh agus an oidhirp reusanachaidh. Cleachdaidh e modail open-weight hosted, a chuireas an sàs iad uile.

from openai import OpenAI

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

response = client.chat.completions.create(
    model="DeepSeek-V4-Flash-0731-W4A16-AUTOROUND-REAP",
    messages=[
        {"role": "system", "content": "You are a physics teacher. Answer in two sentences."},
        {"role": "user", "content": "Why is the sky blue?"},
    ],
    max_tokens=512,
    temperature=0.3,
    top_p=0.9,
    seed=7,
    stop=["\n\n"],
    reasoning_effort="low",
)

message = response.choices[0].message
print(message.reasoning_content)  # the reasoning
print(message.content)            # the answer
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.chat.completions.create({
  model: "DeepSeek-V4-Flash-0731-W4A16-AUTOROUND-REAP",
  messages: [
    { role: "system", content: "You are a physics teacher. Answer in two sentences." },
    { role: "user", content: "Why is the sky blue?" },
  ],
  max_tokens: 512,
  temperature: 0.3,
  top_p: 0.9,
  seed: 7,
  stop: ["\n\n"],
  reasoning_effort: "low",
});

const message = response.choices[0].message;
console.log(message.reasoning_content); // the reasoning
console.log(message.content);           // the answer
console.log(response.usage);

curl https://api.shannon-ai.com/v1/chat/completions \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "DeepSeek-V4-Flash-0731-W4A16-AUTOROUND-REAP",
    "messages": [
      {"role": "system", "content": "You are a physics teacher. Answer in two sentences."},
      {"role": "user", "content": "Why is the sky blue?"}
    ],
    "max_tokens": 512,
    "temperature": 0.3,
    "top_p": 0.9,
    "seed": 7,
    "stop": ["\n\n"],
    "reasoning_effort": "low"
  }'

Tha an aon chruth air a’ fhreagairt ri gu h-àrd. Cuiridh an usage aige dà fhiosrachadh ris air na modailean open-weight hosted: na prompt tokens a chaidh a leughadh on cache agus na tokens a chaidh a chaitheamh air reusanachadh.

{
  "usage": {
    "prompt_tokens": 31,
    "completion_tokens": 62,
    "total_tokens": 93,
    "prompt_tokens_details": {
      "cached_tokens": 0
    },
    "completion_tokens_details": {
      "reasoning_tokens": 21
    }
  }
}

Fad an output

Tha dà rud aig max_tokens. An toiseach, is e an àireamh de tokens a thèid a chur an dàrna taobh bhon balance agad nuair a thòisicheas an t-iarrtas. Nuair a bhios am freagairt coileanta, thèid na tokens a chleachd an t-iarrtas a chur na àite. Ma tha max_tokens nas motha na tha air fhàgail dhen balance agad, tillidh an t-iarrtas 429 Quota exceeded fiù ‘s ged a bhiodh am freagairt fhèin air lùbadh a-steach. Cuir max_tokens nas lugha gus nas lugha a chur an dàrna taobh.

Thèid shannon-coder-1 a chunntadh ann an dòigh eadar-dhealaichte air an endpoint seo: tha gach iarrtas na aon de ghairmean Shannon Coder a’ phlana agad, agus chan eil tokens air an cur an dàrna taobh air a shon. Crìochan agus balance

San dàrna àite, cuingichidh e fad an fhreagairt air na modailean seo:

Às aonais max_tokens no max_completion_tokens, is e 4,096 an luach. Air shannon-coder-1 is e 65,536.

Teachdaireachdan

Tha gach teachdaireachd na nì le role agus content. Is e content sreang, no array de phàirtean nuair a tha barrachd na teacsa san teachdaireachd.

Le id à teaghlach Shannon 3, cuir na stiùiridhean a dh’fheumas a bhith ann dhan teachdaireachd user.

Air ìrean Shannon tillidh iarrtas gun teacsa neach-cleachdaidh agus gun tools 400 No user message provided.

Pàirtean susbaint

Tha duilleag fhèin aig meudan, crìochan agus an liosta slàn de chruthan. Dealbhan agus faidhlichean

Nì an fhreagairt

Adhbharan crìochnachaidh

Bheir freagairt gun streaming cunntas air stop no tool_calls.

Usage

Air na modailean open-weight hosted is e prompt_tokens na teachdaireachdan agus mìneachaidhean inneil agad air an cunntadh le tokenizer a’ mhodail fhèin, a bharrachd air tokens dhealbhan sam bith. Tillidh na h-endpoints cunntais token an aon àireamh mus cuir thu. Cunntadh tokens

Air ìrean Shannon, cunntaidh prompt_tokens a h-uile càil a leugh am modail gus am freagairt a sgrìobhadh, agus mar sin tha e nas motha na teacsa nan teachdaireachdan agad a-mhàin.

Streaming

Le stream air a shuidheachadh gu true thig am freagairt mar thachartasan chat.completion.chunk agus thig e gu crìch le data: [DONE]. Tha finish_reason agus usage aig an chunk mu dheireadh ron sin; chan eil feum air stream_options. Tha duilleag fhèin aig cruthan nan chunks, loidhnichean keep-alive agus mearachdan am broinn stream. Sruthadh

Mearachdan

Is e nì JSON le ball error a th’ ann am mearachd. Ruithidh na sgrùdaidhean san òrdugh seo: iuchair API, bodhaig an iarrtais, id modail, an uair sin balance. Tha an clàr a’ liostadh na tha an endpoint seo a’ tilleadh as trice. Tha duilleag fhèin aig an liosta slàn, le na ghabhas feuchainn a-rithist. Làimhseachadh mearachd

{
  "error": {
    "type": "invalid_request_error",
    "message": "unknown model: no-such-model"
  }
}