Baltor Get started

Models

Hy4 preview

By Tencent. 780.0 billion parameters, a context window of 1,048,576 tokens and the licence apache-2.0.

Facts

Released
2026-08-27 the day the repository was first published on Hugging Face Hugging Face, read
Licence
apache-2.0 the licence the model card declares Hugging Face, read
Open weights
Yes the weights are published in this Hugging Face repository Hugging Face, read
Parameters
780.0 billion counted from the safetensors weight files Hugging Face, read
Active parameters
49.7 billion estimate estimated from the expert counts and sizes in the configuration model configuration, read
Knowledge cutoff
Unknown
Context window
  • 1,048,576 tokens the maximum position embeddings in the model configuration model configuration, read
  • 1,048,576 tokens the context length OpenRouter lists OpenRouter, read
Longest output
  • 64,000 tokens the largest output OpenRouter's first provider allows OpenRouter, read
Tool calling
  • Yes OpenRouter lists tools among the supported parameters OpenRouter, read
Structured output
  • Yes OpenRouter lists structured outputs among the supported parameters OpenRouter, read
Reasoning controls
  • Yes OpenRouter lists reasoning controls, with the efforts high, low, none OpenRouter, read
Inputs and outputs
text in, text out Hugging Face, read
Good for
  • Reasoning: OpenRouter lists reasoning controls for it OpenRouter, read

Prices

Prices in US dollars per million tokens, as each source lists them
ProviderInputOutputContextAs ofSource
Hugging Face direct0.834 USD2.50 USD1,000,000models.dev, read
Kilo Gateway direct0.834 USD2.50 USD1,048,576models.dev, read
LLM Gateway direct0.834 USD2.50 USD1,000,000models.dev, read
NanoGPT direct0.834 USD2.50 USD1,048,576models.dev, read
Novita through OpenRouter0.834 USD2.50 USD1,000,000OpenRouter endpoints, read
OpenRouter direct0.834 USD2.50 USD1,048,576models.dev, read
SiliconFlow through OpenRouter0.834 USD2.50 USD1,048,576OpenRouter endpoints, read
Tencent through OpenRouter0.834 USD2.50 USD1,048,576OpenRouter endpoints, read
Vercel AI Gateway direct0.834 USD2.50 USD1,024,000models.dev, read
CrossModel direct0.960 USD2.88 USD1,048,576models.dev, read

Run it on your own hardware

Weights are the sizes of the files a source lists, or an estimate from the parameter count where none does. Memory is an estimate: weights plus KV cache plus 512 MiB and 5 percent of the weights for runtime buffers. Check it against your hardware.

QuantizationWeightsMemory at 8,192 tokensMemory at 32,768 tokens
IQ2_XS219 GiB231 GiB233 GiB
Q2_K270 GiB284 GiB286 GiB
IQ4_XS388 GiB409 GiB411 GiB

Start it with a local runtime

  • Ollama
    ollama run hf.co/6block/Hy4-preview-GGUF:IQ2_XS
  • LM Studio
    lms get 6block/Hy4-preview-GGUF
lms server start
  • llama.cpp server
    llama-server -hf 6block/Hy4-preview-GGUF:IQ2_XS --jinja
  • vLLM
    vllm serve tencent/Hy4-preview
  • SGLang
    sglang serve --model-path tencent/Hy4-preview --port 30000

Use it from your harness

Set up for CrossModel with the model tencent/hy4-preview. Each endpoint page has the same setup for its own address.

OpenCode

Put this in opencode.json in your project folder:

{
  "$schema": "https://opencode.ai/config.json",
  "provider": {
    "crossmodel": {
      "npm": "@ai-sdk/openai-compatible",
      "name": "CrossModel",
      "options": {
        "baseURL": "https://api.crossmodel.ai/v1",
        "apiKey": "{env:CROSSMODEL_API_KEY}"
      },
      "models": {
        "tencent/hy4-preview": {
          "name": "Hy4 preview",
          "limit": {
            "context": 1048576,
            "output": 64000
          }
        }
      }
    }
  }
}
  • OpenCode reads any OpenAI-compatible address through the @ai-sdk/openai-compatible package, and an address that speaks the Responses API through @ai-sdk/openai.

From OpenCode documentation, read .

Pi

Put this in ~/.pi/agent/models.json:

{
  "providers": {
    "crossmodel": {
      "baseUrl": "https://api.crossmodel.ai/v1",
      "api": "openai-completions",
      "apiKey": "$CROSSMODEL_API_KEY",
      "models": [
        {
          "id": "tencent/hy4-preview"
        }
      ]
    }
  }
}
  • The apiKey field can name an environment variable as $NAME.

From Pi documentation, read .

Codex

Codex speaks only the Responses API, and CrossModel documents no Responses address. A gateway that offers one can sit in between.

  • Codex speaks the Responses API only: responses is the one supported wire API of a custom provider. Ollama and LM Studio are built in and start with --oss.

From Codex documentation, read .

Claude Code

Claude Code sends Anthropic Messages requests, and CrossModel documents no such address. A gateway that translates to that API can sit in between.

  • Claude Code sends Anthropic Messages requests to ANTHROPIC_BASE_URL. Anthropic says it does not support routing Claude Code to models other than Claude through any gateway, so some features may not work with another model.

From Claude Code documentation, read .

Other ways to reach it

Published results

These are results other people published. Baltor did not run them and does not rank models by them.