Chat Completions

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

Isicelo esincane kakhulu siyi-id ye-model nomlayezo owodwa womsebenzisi.

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."}]
  }'

Impendulo iyinto eyodwa ye-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
  }
}

Ama-header

Ama-header esicelo

Ama-header empendulo

Ama-field esicelo

I-messages kuphela edingekayo. Ikholomu ethi Kusebenza ku- ibala ama-model lapho i-field ishintsha khona impendulo. Ama-model e-open-weight ahlinzekiwe ama-id ayishumi nambili osukwini lwama-model; umndeni we-Shannon 3 ngu-shannon-3, shannon-3-pro, shannon-3.1 no-shannon-3.1-pro. Ama-model namanani

Ezinye ama-field e-OpenAI, afana ne-n, user, stream_options, parallel_tool_calls, presence_penalty, frequency_penalty, logit_bias, logprobs, metadata, store ne-prompt_cache_key, ayamukelwa ukuze ikhodi yeklayenti esekhona isebenze ngaphandle kokushintshwa. Awashintshi impendulo: njalo kunokukhetha okukodwa, futhi i-stream iphela njalo ngokusetshenziswa.

I-field enohlobo olungalungile lwe-JSON, isibonelo "max_tokens": "100", ibuyisela i-422. Isicelo esingenayo i-messages senza okufanayo.

Amathuluzi, okuphumayo okuhleliwe, ukucabanga nokusesha kuwebhu kunekhasi lakho ngalinye: Ukubiza umsebenzi, Imiphumela ehlanganisiwe, Umzamo wokucabanga, Usesho lwewebhu.

Isicelo esinezinketho

Lesi sicelo sisetha umlayezo wesistimu, ama-field e-sampling nomzamo wokucabanga. Sisebenzisa i-model ye-open-weight ehlinzekiwe, esebenzisa zonke lezi.

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"
  }'

Impendulo inesimo esifanayo nesingenhla. I-usage yayo yengeza imininingwane emibili kuma-model e-open-weight ahlinzekiwe: ama-token e-prompt afundwe ku-cache nama-token asetshenziselwe ukucabanga.

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

Ubude bokuphumayo

I-max_tokens yenza izinto ezimbili. Okokuqala, inani lama-token abekelwa eceleni ku-balance yakho lapho isicelo siqala. Uma impendulo isiphelele, lelo nani lifakwa esikhundleni sama-token isicelo esizisebenzisile. Uma i-max_tokens inkulu kunalokho okusele ku-balance yakho, isicelo sibuyisela i-429 Quota exceeded noma impendulo ngokwayo ibizofaneleka. Thumela i-max_tokens encane ukuze ubeke eceleni okuncane.

I-shannon-coder-1 ibalwa ngokuhlukile kule endpoint: isicelo ngasinye siyisicelo esisodwa se-Shannon Coder sepulani yakho, futhi akukho ma-token abekelwa eceleni. Imikhawulo ne-balance

Okwesibili, ikhawulela ubude bempendulo kulawa ma-model:

Ngaphandle kwe-max_tokens noma i-max_completion_tokens, inani lingu-4,096. Ku-shannon-coder-1 lingu-65,536.

Imilayezo

Umlayezo ngamunye uyinto ene-role ne-content. I-content iyi-string, noma uhlu lwezingxenye uma umlayezo uphatha okungaphezu kombhalo.

Nge-id yomndeni we-Shannon 3, faka imiyalo okufanele ilandelwe emlayezweni we-user.

Emazingeni e-Shannon isicelo esingenawo umbhalo womsebenzisi futhi singenawo ama-tools sibuyisela i-400 No user message provided.

Izingxenye zokuqukethwe

Osayizi, imikhawulo nohlu olugcwele lwamafomu zinekhasi lazo. Izithombe namafayela

Into yempendulo

Izizathu zokuphela

Impendulo engasakazwa ibika i-stop noma i-tool_calls.

Ukusetshenziswa

Kuma-model e-open-weight ahlinzekiwe, i-prompt_tokens yimilayezo yakho nezincazelo zamathuluzi zibalwa nge-tokenizer ye-model ngokwayo, kanye nama-token ezithombe zonke. Ama-endpoint okubala ama-token abuyisela inombolo efanayo ngaphambi kokuthumela. Ukubala ama-token

Emazingeni e-Shannon, i-prompt_tokens ibala konke i-model eyakufunda ukubhala impendulo, ngakho inkulu kunombhalo wemilayezo yakho wodwa.

I-Streaming

Nge-stream esethelwe ku-true impendulo ifika njengemicimbi ye-chat.completion.chunk bese iphela nge-data: [DONE]. I-chunk yokugcina ngaphambi kwayo iphatha i-finish_reason ne-usage; akudingeki ama-stream_options. Izimo ze-chunk, imigqa ye-keep-alive namaphutha ngaphakathi kwe-stream kunekhasi lakho. Ukusakaza

Amaphutha

Iphutha liyinto ye-JSON enelungu le-error. Ukuhlola kwenziwa ngalolu hlelo: i-API key, umzimba wesicelo, i-id ye-model, bese i-balance. Ithebula libala lokho le-endpoint ebuyisela kaningi. Uhlu olugcwele, nalokho okufanele kuzanywe futhi, lunekhasi lalo. Ukuphathwa kwamaphutha

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