Chat Completions

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

Ko te tono iti rawa he model id me tētahi karere kaiwhakamahi kotahi.

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."}]
  }'

He JSON object kotahi te whakautu:

{
  "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
  }
}

Ngā header

Ngā header tono

Ngā header whakautu

Ngā āpure tono

Ko te messages anake te mea e hiahiatia ana. Ko te pou Whakamahia e e ingoa ana i ngā model e huri ai te whakautu mā tētahi āpure. Ko ngā model open-weight hosted ko 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

Ka whakaaetia ētahi atu āpure OpenAI, pērā i te n, user, stream_options, parallel_tool_calls, presence_penalty, frequency_penalty, logit_bias, logprobs, metadata, store me prompt_cache_key, kia kore ai e huri te code kiritaki o nāianei. Kāore ō rātou pānga ki te whakautu: kotahi tonu te choice, ā, ka mutu tonu tētahi stream me te usage.

Ka whakahoki tētahi āpure he hē te momo JSON, hei tauira "max_tokens": "100", i te 422. Pērā hoki te tono kāore he messages.

He whārangi motuhake tō ngā tools, te output whakatakoto, te reasoning me te rapu tukutuku: Karanga ā‑mahi, Putanga hanganga, Reasoning effort, Rapu paetukutuku.

He tono me ngā kōwhiringa

Ka tautuhi tēnei tono i tētahi karere system, i ngā āpure sampling me te reasoning effort. Ka whakamahi i tētahi model open-weight hosted, ka whakamahi i te katoa.

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"
  }'

He āhua ōrite te whakautu ki runga ake nei. Ka tāpiri tōna usage e rua ngā taipitopito i runga i ngā model open-weight hosted: ngā prompt tokens i pānuitia mai i te cache me ngā tokens i whakapaua ki te reasoning.

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

Te roa o te output

E rua ngā mahi a te max_tokens. Tuatahi, ko te maha o ngā tokens ka wehea mai i tō toenga ina tīmata te tono. Ina oti te whakautu, ka whakakapia taua nui e ngā tokens i whakamahia e te tono. Mēnā he nui ake te max_tokens i te toenga o tō toenga, ka whakahoki te tono i te 429 Quota exceeded ahakoa ka tau te whakautu ake. Tukuna he max_tokens iti ake kia iti iho te wehenga.

Ka tatauria te shannon-coder-1 i tētahi tikanga rerekē i runga i tēnei endpoint: he karanga Shannon Coder o tō mahere te tono kotahi, ā, kāore he tokens e wehea mō taua tono. Ngā tepe me te toenga

Tuarua, ka whakawhāiti i te roa o te whakautu i runga i ēnei model:

Mēnā kāore he max_tokens he max_completion_tokens rānei, ko te uara 4,096. I runga i te shannon-coder-1 ko 65,536.

Ngā karere

He object ia karere me tētahi role me tētahi content. Ko te content he rārangi tuhi, he rārangi wāhanga rānei ina nui ake te kawe a te karere i te kuputuhi.

Me tētahi id o te whānau Shannon 3, whakatōngia ngā tohutohu me mau ki roto i te karere user.

I runga i ngā taumata Shannon ka whakahoki te tono kāore he kuputuhi kaiwhakamahi, kāore he tools i te 400 No user message provided.

Ngā wāhanga ihirangi

He whārangi motuhake tō ngā rahi, ngā tepe me te rārangi katoa o ngā āhua. Ngā whakaahua me ngā kōnae

Te object whakautu

Ngā take mutunga

Ka pūrongo te whakautu kāore i stream i te stop he tool_calls rānei.

Usage

I runga i ngā model open-weight hosted, ko te prompt_tokens ō karere me ō tautuhinga tool kua tatauria ki te tokenizer ake a te model, tāpiri ngā tokens o ngā whakaahua. Ka whakahoki ngā endpoint tatau token i te tau ōrite i mua i tō tuku. Te tatau token

I runga i ngā taumata Shannon, ka tatau te prompt_tokens i ngā mea katoa i pānuitia e te model hei tuhi i te whakautu, nō reira he nui ake i te kuputuhi o ō karere anake.

Streaming

Ina tautuhia te stream ki te true ka tae mai te whakautu hei tuhinga chat.completion.chunk ka mutu me te data: [DONE]. Ka kawe te wāhanga whakamutunga i mua i a ia i te finish_reason me te usage; kāore he stream_options e hiahiatia. He whārangi motuhake tō ngā āhua wāhanga, ngā rārangi keep-alive me ngā hapa i roto i tētahi stream. Rere

Ngā hapa

He JSON object te hapa me tētahi mema error. Ka whakahaerea ngā tirotiro i tēnei raupapa: API key, tinana tono, model id, kātahi te toenga. Ka rārangi te ripanga i ngā mea e hoki mai ana i tēnei endpoint i te nuinga o te wā. He whārangi motuhake tō te rārangi katoa, me ngā mea hei whakamātau anō. Whakahaere hapa

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