Description

A sophisticated text-based Mixture-of-Experts (MoE) model featuring 21B total parameters with 3B activated per token, delivering exceptional multimodal understanding and generation through heterogeneous MoE structures and modality-isolated routing. Supporting an extensive 131K token context length, the model achieves efficient inference via multi-expert parallel collaboration and quantization, while advanced post-training techniques including SFT, DPO, and UPO ensure optimized performance across diverse applications with specialized routing and balancing losses for superior task handling.

Technical Specifications

Example API Call

curl -X POST https://api.neuralhub.xyz/v1/chat/completions \
-H "Content-Type: application/json" \
-H "Authorization: Bearer NEURALHUB_API_KEY" \
-d '{
  "model": "baidu/ernie-4.5-21b-a3b",
  "messages": [
    { "role": "system", "content": "You are a helpful assistant." },
    { "role": "user", "content": "" }
  ],
  "temperature": 0.7,
  "max_tokens": 500,
  "top_p": 0.9
}'

Response Format

The API returns an OpenAI-compatible response. Example:

{
  "id": "chatcmpl-<uuid>",
  "object": "chat.completion",
  "created": 1765590402,
  "model": "baidu/ernie-4.5-21b-a3b",
  "choices": [
    {
      "index": 0,
      "message": {
        "role": "assistant",
        "content": "The answer to life, the universe, and everything is famously 42..."
      },
      "finish_reason": "stop"
    }
  ],
  "usage": {
    "prompt_tokens": 26,
    "completion_tokens": 169,
    "total_tokens": 195
  }
}