Modella et pretia

modella sine censura

22

fenestra contextus

256K

a

$0.50 /1M

Modella open-weight hospitata

contextus 262,144 tokenorum in omni id; input et output separatim computantur; quodque id accipit quod modellum quod imitatur accipit

Modellum	Input	Exitus structus	In / cache / ex per 1M	Plura
DeepSeek-V4-Pro-08133BIT-REAPDeepSeek · 1.6T MoE · 49B activeSine censura	Textus in	Schema JSON	$1.95 / $0.488 / $3.90	Plura
GLM-5.23BIT-REAPZ.ai · 744B MoE · 40B activeSine censura	Textus in	Schema JSON	$0.73 / $0.183 / $2.34	Plura
Kimi-K33BIT-REAPMoonshot AI · 2.8T MoE · 104B activeSine censura	Textus + imago in	Schema JSON	$3.83 / $0.958 / $19.12	Plura
Nemotron3Ultra3BIT-REAPNVIDIA · 550B hybrid Mamba-Attention MoE · 55B activeSine censura	Textus in	Schema JSON	$0.75 / $0.188 / $3.30	Plura
MiniMax-M33BIT-REAPMiniMax · 428B MoE · 23B active · sparse attentionSine censura	Textus + imago in	Schema JSON	$0.50 / $0.125 / $2.00	Plura
DeepSeek-V4-Flash-0731W4A16-AUTOROUND-REAPDeepSeek · 284B MoE · 13B activeSine censura	Textus in	Schema JSON	$0.50 / $0.125 / $2.00	Plura
Kimi-K2.6W4A16-AUTOROUND-REAPMoonshot AI · 1T MoE · 32B activeSine censura	Textus + imago in	Schema JSON	$0.78 / $0.195 / $3.67	Plura
Laguna-S-2.1W4A16-AUTOROUND-REAPPoolside · 118B MoE · 8B activeSine censura	Textus in	Sine exitu structo	$0.50 / $0.125 / $2.00	Plura
inklingW4A16-AUTOROUND-REAPThinking Machines · 975B MoE · 41B activeSine censura	Textus + imago in	Objectum JSON	$1.42 / $0.355 / $6.07	Plura
MiMo-V2.5-ProW8A16Xiaomi · 1.02T MoE · 42B active · 8-bitSine censura	Textus in	Schema JSON	$0.50 / $0.125 / $2.00	Plura
MiMo-V2.5W8A16Xiaomi · 8-bit multimodal · W8A16Sine censura	Textus + imago in	Schema JSON	$0.50 / $0.125 / $2.00	Plura
Hy3W8A16Tencent · 295B MoE · 21B active · 8-bitSine censura	Textus in	Schema JSON	$0.50 / $0.125 / $2.00	Plura

Modella Shannon

una merces plana pro input et output

Modellum	Contextus	Input	Exitus structus	Pretium / 1M tokens
Shannon 1.6 Liteshannon-1.6-liteSine censura	192K	Textus + imago in	Schema JSON	$3.90
Shannon 1.6 Proshannon-1.6-proSine censura	192K	Textus + imago in	Schema JSON	$7.80
Shannon 2 Liteshannon-2-liteSine censura	192K	Textus + imago in	Schema JSON	$3.90
Shannon 2 Proshannon-2-proSine censura	192K	Textus + imago in	Schema JSON	$5.85
Shannon 3shannon-3Sine censura	192K	Textus + imago in	Schema JSON	$3.35
Shannon 3 Proshannon-3-proSine censura	192K	Textus + imago in	Schema JSON	$3.35
Shannon 3.1shannon-3.1Sine censura	192K	Textus + imago in	Schema JSON	$3.35
Shannon 3.1 Proshannon-3.1-proSine censura	192K	Textus + imago in	Schema JSON	$3.35
Shannon 3.5shannon-3.5Sine censura	192K	Textus + imago in	Schema JSON	$3.35
Shannon Coder 1shannon-coder-1Sine censura	128K	Textus in	Schema JSON	$8.00

Ad Initium celere salire Playground experire

Quomodo lineam legas

Omne modellum streaming praestat et tools vocare potest. Aliae facultates API in pagina Capabilities describuntur. Facultates

Context window et limites outputi

• Context window omnia comprehendit quae modellum pro una responsione legit: nuntios tuos, definitiones tool et imagines.
• Familia Shannon 3 conversationem longiorem per se aptat: turnos novissimos qui in fenestram capiuntur servat et respondet.
• In stream, prompt nimis longum nuntiatur ut error frame cum type invalid_request_error et code: "context_length_exceeded".
• max_tokens valores ab 1 ad 65,536 in omni modello accipit; max_output_tokens in GET /v1/models est summum illius range. Defaultum est 4,096.
• In omni modello, max_tokens est quantitas quam petitio de saldo tuo reservat.
• Ut limes longitudinis responsionis, max_tokens a modellis open-weight hosted, shannon-1.6-lite, shannon-1.6-pro et shannon-coder-1 applicatur.
• In modellis open-weight hosted textum responsionis limitat. Ratiocinatio ante responsionem contra eum non computatur, et valor infra 256 ut 256 agit.
• Stream shannon-1.6-lite, shannon-1.6-pro, shannon-coder-1 vel familiae Shannon 3 cum finish_reason: "length" finit cum responsio ad limitem suum abscisa est.

Limites et saldum

Quomodo pretium formetur

• Pretia sunt in USD per 1,000,000 tokenum. Saldum tuum in tokens computatur qui $5.00 per 1,000,000 valent, itaque pretium modelli dicit quantum saldi tokens eius capiant.
• Modelli Shannon unum pretium habent. Vocatio computatur usage.total_tokens ad illud pretium.
• Modelli open-weight hosted tria pretia habent. Input e prompt cache lectum ad pretium cached computatur, quod est 25% pretii input ad $0.001 rotundati. Reliquum inputi ad pretium input computatur, output ad pretium output.
• Modelli open-weight hosted tokens cum tokenizer proprio cuiusque modelli computant. GET /v1/models hoc ut tokenizer: "original" notat.
• shannon-coder-1 in vocationibus computatur, non in tokens, in /v1/chat/completions et /v1/messages. Pretium eius indicatum in /v1/responses valet.

Quid petitio reservet, exempla computata pretii et quota vocationum Shannon Coder in Limits and balance sunt. Limites et saldum

Ids modellorum

Mitte unum ex his ids ut model.

shannon-1.6-lite
shannon-1.6-pro
shannon-2-lite
shannon-2-pro
shannon-3
shannon-3-pro
shannon-3.1
shannon-3.1-pro
shannon-3.5
shannon-coder-1
DeepSeek-V4-Pro-0813-3BIT-REAP
GLM-5.2-3BIT-REAP
Kimi-K3-3BIT-REAP
Nemotron3Ultra-3BIT-REAP
MiniMax-M3-3BIT-REAP
DeepSeek-V4-Flash-0731-W4A16-AUTOROUND-REAP
Kimi-K2.6-W4A16-AUTOROUND-REAP
Laguna-S-2.1-W4A16-AUTOROUND-REAP
inkling-W4A16-AUTOROUND-REAP
MiMo-V2.5-Pro-W8A16
MiMo-V2.5-W8A16
Hy3-W8A16

• Litterae maiusculae et minusculae nihil referunt, et spatia circa id removentur: SHANNON-3 et shannon-3 idem modellum sunt.
• Responsio semper id publicatum in model fert, quamcumque scripturam misisti.
• Nomen ostensionis, ut Shannon 3, non est id.

Quattuor scripturae breviores etiam accipiuntur.

Id quod non est publicatum respondetur cum statu 400, type invalid_request_error et nuntio unknown model: <id> in omnibus tribus endpoints. Nullum aliud modellum loco eius respondet.

{
  "error": {
    "type": "invalid_request_error",
    "message": "unknown model: gpt-4o"
  }
}

{
  "type": "error",
  "error": {
    "type": "invalid_request_error",
    "message": "unknown model: gpt-4o"
  }
}

• Semper mitte model. /v1/responses petitioni sine eo cum 400 respondet. In /v1/chat/completions et /v1/messages talis petitio a shannon-1.6-lite respondetur.
• POST /v1/tokenize et POST /v1/messages/count_tokens ids modellorum open-weight hosted accipiunt.

Modellos enumerare

GET /v1/models modellos huius paginae ut JSON reddit, cum iisdem context windows, pretiis et facultatibus. Utere eo ad model picker implendum vel ad pretium in codice verificandum.

GET https://api.shannon-ai.com/v1/models

from openai import OpenAI

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

for model in client.models.list():
    print(model.id)

import OpenAI from 'openai';

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

for await (const model of client.models.list()) {
  console.log(model.id);
}

curl "https://api.shannon-ai.com/v1/models"

Responsio, ad unam inscriptionem cuiusque generis breviata:

{
  "object": "list",
  "data": [
    {
      "id": "shannon-3",
      "object": "model",
      "created": 1700000000,
      "owned_by": "shannon-ai",
      "display_name": "Shannon 3",
      "context_window": 196608,
      "max_output_tokens": 65536,
      "pricing": {
        "input_per_million_usd": 3.35,
        "output_per_million_usd": 3.35,
        "currency": "USD"
      },
      "capabilities": {
        "tools": true,
        "vision": true,
        "reasoning": true,
        "reasoning_effort": false,
        "streaming": true,
        "json_schema": true
      },
      "endpoints": [
        "/v1/chat/completions",
        "/v1/messages",
        "/v1/responses"
      ]
    },
    {
      "id": "MiniMax-M3-3BIT-REAP",
      "object": "model",
      "created": 1700000000,
      "owned_by": "shannon-ai",
      "display_name": "MiniMax-M3 · 3BIT-REAP",
      "context_window": 262144,
      "max_output_tokens": 65536,
      "pricing": {
        "input_per_million_usd": 0.5,
        "cached_input_per_million_usd": 0.125,
        "output_per_million_usd": 2,
        "currency": "USD"
      },
      "tokenizer": "original",
      "capabilities": {
        "tools": true,
        "vision": true,
        "reasoning": true,
        "reasoning_effort": true,
        "streaming": true,
        "response_format": true,
        "json_schema": true,
        "prompt_caching": true
      },
      "endpoints": [
        "/v1/chat/completions",
        "/v1/messages",
        "/v1/responses",
        "/v1/messages/count_tokens",
        "/v1/tokenize"
      ]
    }
  ]
}

• GET /v1/models nullam clavem requirit. Responsio eadem est pro omni vocante nec a plano tuo pendet.
• Index ids quae vocari possunt tenet. Modellum quod hoc momento servari non potest ab eo omittitur.
• Nullum endpoint est pro uno modello: GET /v1/models/{id} respondet 404 cum type not_found_error.