Models

Get started
Cognitivess Cognitivess Models
Cognitivess-1
Flagship model developed by CognitivessAI.
Model ID Cognitivess-1
Context length1,048,576 tokens
Max output131,072 tokens
DeveloperCognitivessAI
ProtocolOpenAI-compatible · Anthropic-compatible
Input / Outputtext → text
Pricing (per 1M tokens, pay only for usage)
Input
$1.96
Cache read
$0.36
Output
$6.16
Cloud External Models
Kling v3.0 4K Text-to-Video
Kling v3.0 4K video generation.
Model ID kling-v3.0-4k-t2v
TypeVideo
DeveloperKling
Resolution4K
AudioNo Audio
ProtocolCognitivessAI async video API
Input / Outputtext → video
Pricing (pay only for usage)
Per second
$0.59 /s

View API reference →

Kling v3.0 4K Image-to-Video
Kling v3.0 4K video generation.
Model ID kling-v3.0-4k-i2v
TypeVideo
DeveloperKling
Resolution4K
AudioNo Audio
ProtocolCognitivessAI async video API
Input / Outputimage → video
Pricing (pay only for usage)
Per second
$0.59 /s

View API reference →

Local Local Models — run on your machine
Open-weight GGUF models you can run fully offline with the cognitivess CLI — no API key, no billing. Pull one with cognitivess pull <name>, then serve it in a browser with cognitivess serve <name> or run it inside a coding tool with cognitivess launch claude --local <name>. CLI reference →
qwen2.5-0.5b
Tiny instruct model (0.5B) — fast, runs anywhere. Good for testing the local pipeline.
SourceQwen/Qwen2.5-0.5B-Instruct-GGUF
Size~0.4 GB
Quantqwen2.5-0.5b-instruct-q4_k_m.gguf
cognitivess pull qwen2.5-0.5b
llama3
Llama 3.1 8B Instruct — general-purpose chat & light coding.
Sourceunsloth/llama-3.1-8b-instruct-gguf
Size~4.9 GB
Quantllama-3.1-8b-instruct-Q4_K_M.gguf
cognitivess pull llama3
qwen2.5-7b
Qwen2.5 7B Instruct — strong general chat & multilingual.
SourceQwen/Qwen2.5-7B-Instruct-GGUF
Size~4.7 GB
Quantqwen2.5-7b-instruct-q4_k_m.gguf
cognitivess pull qwen2.5-7b
glm-ocr
Vision OCR model — extracts text from images & documents.
vision
Sourceggml-org/GLM-OCR-GGUF
Size~0.9 GB
QuantGLM-OCR-Q8_0.gguf
cognitivess pull glm-ocr
laguna-xs-2.1
Agentic coding & long-horizon work — 33B MoE (3B active), 256K context. Note: needs a llama.cpp build from PR #25165 (not the auto-downloaded engine).
Sourcepoolside/Laguna-XS-2.1-GGUF
Size~18.9 GB
QuantLaguna-XS-2.1-Q4_K_M.gguf
cognitivess pull laguna-xs-2.1
ornith-35b
Agentic coding — tool-calling & terminal agents. 35B MoE, 256K context, structured reasoning.
Sourcedeepreinforce-ai/Ornith-1.0-35B-GGUF
Size~19.7 GB
Quantornith-1.0-35b-Q4_K_M.gguf
cognitivess pull ornith-35b
qwen3.6-35b-a3b
Multimodal MoE — agentic coding + vision (images/video/OCR). 35B total / 3B active, up to 1M context.
vision
Sourceunsloth/Qwen3.6-35B-A3B-GGUF
Size~20.6 GB
QuantQwen3.6-35B-A3B-UD-Q4_K_M.gguf
cognitivess pull qwen3.6-35b-a3b
north-mini-code
Agentic coding model (Cohere Labs North-Mini-Code). MoE, code-focused.
Sourceunsloth/North-Mini-Code-1.0-GGUF
Size~17.9 GB
QuantNorth-Mini-Code-1.0-UD-Q4_K_M.gguf
cognitivess pull north-mini-code
nemotron-3-nano
NVIDIA Nemotron 3 Nano — reasoning + multilingual + agentic coding. 30B total / 3B active MoE.
Sourceunsloth/Nemotron-3-Nano-30B-A3B-GGUF
Size~22.9 GB
QuantNemotron-3-Nano-30B-A3B-Q4_K_M.gguf
cognitivess pull nemotron-3-nano
glm-4.7-flash
Z.ai GLM-4.7 Flash — fast general + tool-calling. Set --repeat-penalty 1.0 for best results.
Sourceunsloth/GLM-4.7-Flash-GGUF
Size~17.1 GB
QuantGLM-4.7-Flash-Q4_K_M.gguf
cognitivess pull glm-4.7-flash
qwythos-9b
Reasoning model on a Qwen3.5-9B base, post-trained on Claude Mythos/Fable traces. 1M context (YaRN), native function calling, vision. Intentionally uncensored.
vision
Sourceempero-ai/Qwythos-9B-Claude-Mythos-5-1M-GGUF
Size~5.2 GB
QuantQwythos-9B-Claude-Mythos-5-1M-Q4_K_M.gguf
cognitivess pull qwythos-9b

How to use — Cognitivess-1

Point the OpenAI client at CognitivessAI and request Cognitivess-1.
from openai import OpenAI

client = OpenAI(
    api_key="<YOUR_API_KEY>",
    base_url="https://api.cognitivess.com/v1",
)
resp = client.chat.completions.create(
    model="Cognitivess-1",
    messages=[{"role": "user", "content": "Hello"}],
)
print(resp.choices[0].message.content)
import OpenAI from "openai";

const client = new OpenAI({
  apiKey: "<YOUR_API_KEY>",
  baseURL: "https://api.cognitivess.com/v1",
});
const resp = await client.chat.completions.create({
  model: "Cognitivess-1",
  messages: [{ role: "user", content: "Hello" }],
});
console.log(resp.choices[0].message.content);
curl https://api.cognitivess.com/v1/chat/completions \
  -H "Authorization: Bearer <YOUR_API_KEY>" \
  -H "Content-Type: application/json" \
  -d '{"model": "Cognitivess-1", "messages": [{"role":"user","content":"Hello"}]}'

Generate an API key