Chat Completions

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

Codsiga ugu yar waa id model iyo hal fariin isticmaale.

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."}]
  }'

Jawaabtu waa hal shay JSON ah:

{
  "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
  }
}

Header-yo

Header-yada codsiga

Header-yada jawaabta

Field-yada codsiga

Kaliya messages ayaa loo baahan yahay. Tiirka Waxay dabaqaan wuxuu magacaabayaa model-yada field-ku jawaabta ku beddelo. Model-yada open-weight ee la martigeliyo waa laba iyo toban id oo ka mid ah liiska model-yada; qoyska Shannon 3 waa shannon-3, shannon-3-pro, shannon-3.1 iyo shannon-3.1-pro. Model-yo & qiimo

Field-yada kale ee OpenAI, sida n, user, stream_options, parallel_tool_calls, presence_penalty, frequency_penalty, logit_bias, logprobs, metadata, store iyo prompt_cache_key, waa la aqbalaa si code-ka macmiilka ee jira u shaqeeyo isbeddel la'aan. Jawaabta ma beddelaan: had iyo jeer hal choice ayaa jira, stream-kuna had iyo jeer wuxuu ku dhammaadaa usage.

Field leh nooc JSON qaldan, tusaale ahaan "max_tokens": "100", wuxuu soo celiyaa 422. Codsi aan lahayn messages sidoo kale.

Tools, structured output, reasoning iyo raadinta webka mid kastaa wuxuu leeyahay bog u gaar ah: Wacyigelinta Shaqada, Waxsoosaarka qaabaysan, Effort-ka reasoning-ka, Ku-dhismay Raadinta Shabakadda.

Codsi leh ikhtiyaaro

Codsigani wuxuu dejiyaa fariin system, field-yada sampling iyo reasoning effort. Wuxuu isticmaalaa model open-weight oo la martigeliyay, kaas oo dabaqa dhammaantood.

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"
  }'

Jawaabtu waxay leedahay isla qaabka kor ku xusan. usage kiisu wuxuu ku daraa laba faahfaahin model-yada open-weight ee la martigeliyo: token-yada prompt-ka laga akhriyay cache iyo token-yada loo isticmaalay reasoning.

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

Dhererka output-ka

max_tokens wuxuu sameeyaa laba shay. Marka hore, waa tirada token-yada laga kala dhigo haraaggaaga marka codsigu bilaabmo. Marka jawaabtu dhammaato, cadadkaas waxaa lagu beddelaa token-yada codsigu isticmaalay. Haddii max_tokens ka weyn yahay waxa ka hadhay haraaggaaga, codsigu wuxuu soo celiyaa 429 Quota exceeded xataa haddii jawaabtu laftigeedu soo galeen lahayd. Dir max_tokens hoose si aad wax yar u kala dhigto.

shannon-coder-1 si kale ayaa loo tiriyaa endpoint-kan: codsi kastaa waa mid ka mid ah wicitaannada Shannon Coder ee qorshahaaga, token-na looma kala dhigo. Xadad iyo haraag

Marka labaad, wuxuu xadidaa dhererka jawaabta model-yadan:

max_tokens ama max_completion_tokens la'aan, qiimuhu waa 4,096. shannon-coder-1 waa 65,536.

Fariimaha

Fariin kastaa waa shay leh role iyo content. content waa string, ama array ah qaybo marka fariintu sido wax ka badan qoraal.

Id qoyska Shannon 3 la isticmaalayo, tilmaamaha waajibka ah geli fariinta user.

Heerarka Shannon codsi aan lahayn qoraal isticmaale iyo tools wuxuu soo celiyaa 400 No user message provided.

Qaybaha nuxurka

Cabbirrada, xadadka iyo liiska buuxa ee qaababka waxay leeyihiin bog u gaar ah. Sawirro iyo faylal

Shayga jawaabta

Sababaha dhammaadka

Jawaab aan stream ahayn waxay sheegtaa stop ama tool_calls.

Isticmaal

Model-yada open-weight ee la martigeliyo, prompt_tokens waa fariimahaaga iyo qeexidaha tool oo lagu tiriyay tokenizer-ka model-ka laftiisa, ugu daa token-yada sawirrada oo kale. Endpoint-yada tirinta token-yada waxay soo celiyaan isla tirada ka hor inta aadan dirin. Tirinta token-yada

Heerarka Shannon, prompt_tokens wuxuu tiriyaa wax kasta oo model-ku akhriyay si uu jawaabta u qoro, sidaas darteed wuu ka weyn yahay qoraalka fariimahaaga oo keliya.

Streaming

Marka stream la dejiyo true, jawaabtu waxay timaadaa sida events chat.completion.chunk wuxuuna ku dhammaadaa data: [DONE]. Chunk-ga ugu dambeeya ka hor wuxuu sidaa finish_reason iyo usage; stream_options looma baahna. Qaababka chunk, xariiqaha keep-alive iyo khaladaadka stream dhexdiisa waxay leeyihiin bog u gaar ah. Streaming

Khaladaad

Khaladku waa shay JSON ah oo leh xubin error. Hubinta waxay u socotaa sidan: furaha API, jidhka codsiga, id-ga model-ka, kaddib haraagga. Shaxdu waxay liis garaynaysaa waxa endpoint-kani inta badan soo celiyo. Liiska buuxa, oo ay la socoto waxa dib loo tijaabiyo, wuxuu leeyahay bog u gaar ah. Khaladaadka Maareynta

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