Chat Completions

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

Ombi dogo zaidi ni model id na ujumbe mmoja wa user.

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."}]
  }'

Jibu ni kitu kimoja cha JSON:

{
  "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 za ombi

Headers za jibu

Fields za ombi

messages pekee ndiyo inahitajika. Safu ya Inatumiwa na inataja model ambazo field inabadilisha jibu juu yake. Model za open-weight zinazopangishwa ni id kumi na mbili za orodha ya model; familia ya Shannon 3 ni shannon-3, shannon-3-pro, shannon-3.1 na shannon-3.1-pro. Model na bei

Fields nyingine za OpenAI, kama n, user, stream_options, parallel_tool_calls, presence_penalty, frequency_penalty, logit_bias, logprobs, metadata, store na prompt_cache_key, zinakubaliwa ili code ya client iliyopo ifanye kazi bila mabadiliko. Haziathiri jibu: daima kuna choice moja, na stream daima huisha na usage.

Field yenye aina isiyo sahihi ya JSON, kwa mfano "max_tokens": "100", hurudisha 422. Ombi lisilo na messages pia.

Tools, matokeo yaliyopangwa, reasoning na web search kila moja ina ukurasa wake: Uitoaji wa kazi, Matokeo yaliyopangwa, Reasoning effort, Utafutaji wa wavuti.

Ombi lenye options

Ombi hili linaweka ujumbe wa system, fields za sampling na reasoning effort. Linatumia model ya open-weight inayopangishwa, ambayo huzitumia zote.

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"
  }'

Jibu lina umbo lile lile la juu. usage yake huongeza maelezo mawili kwenye model za open-weight zinazopangishwa: tokens za prompt zilizosomwa kutoka cache na tokens zilizotumika kwa reasoning.

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

Urefu wa output

max_tokens hufanya mambo mawili. Kwanza, ni idadi ya tokens inayotengwa kutoka kwenye salio lako ombi linapoanza. Jibu likikamilika, kiasi hicho hubadilishwa na tokens ambazo ombi lilitumia. Ikiwa max_tokens ni kubwa kuliko kilichobaki kwenye salio lako, ombi hurudisha 429 Quota exceeded hata kama jibu lenyewe lingetosha. Tuma max_tokens ndogo zaidi ili kutenga kidogo.

shannon-coder-1 huhesabiwa tofauti kwenye endpoint hii: kila ombi ni mojawapo ya wito wa Shannon Coder wa mpango wako, na hakuna tokens zinazotengwa kwa ajili yake. Mipaka na salio

Pili, inapunguza urefu wa jibu kwenye model hizi:

Bila max_tokens au max_completion_tokens, thamani ni 4,096. Kwenye shannon-coder-1 ni 65,536.

Ujumbe

Kila ujumbe ni kitu chenye role na content. content ni string, au array ya sehemu wakati ujumbe unabeba zaidi ya maandishi.

Ukiwa na id ya familia ya Shannon 3, weka maagizo yanayopaswa kushikilia kwenye ujumbe wa user.

Kwenye viwango vya Shannon ombi lisilo na maandishi ya user wala tools hurudisha 400 No user message provided.

Sehemu za maudhui

Ukubwa, mipaka na orodha kamili ya maumbo vina ukurasa wake. Picha na mafaili

Kitu cha jibu

Sababu za kumaliza

Jibu lisilo la stream huripoti stop au tool_calls.

Matumizi

Kwenye model za open-weight zinazopangishwa, prompt_tokens ni ujumbe wako na ufafanuzi wa tools uliohesabiwa kwa tokenizer ya model yenyewe, pamoja na tokens za picha zozote. Endpoint za kuhesabu tokens hurudisha namba ile ile kabla ya kutuma. Kuhesabu tokens

Kwenye viwango vya Shannon, prompt_tokens huhesabu kila kitu ambacho model ilisoma kuandika jibu, kwa hiyo ni kubwa kuliko maandishi ya ujumbe wako pekee.

Streaming

Kwa stream iliyowekwa kuwa true jibu hufika kama matukio ya chat.completion.chunk na huisha na data: [DONE]. Chunk ya mwisho kabla yake hubeba finish_reason na usage; stream_options hazihitajiki. Maumbo ya chunk, mistari ya keep-alive na makosa ndani ya stream yana ukurasa wao. Kutiririsha

Makosa

Kosa ni kitu cha JSON chenye member ya error. Ukaguzi hufanyika kwa mpangilio huu: API key, mwili wa ombi, model id, kisha salio. Jedwali linaorodhesha kile ambacho endpoint hii hurudisha mara nyingi zaidi. Orodha kamili, pamoja na yapi ya kujaribu tena, ina ukurasa wake. Ushughulikiaji wa Makosa

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