Dagon Inference / Pricing / Hy4
Hy4 API · Silk Compute
Published · Last updated
Dagon Inference serves Hy4 through an OpenAI compatible API at https://inference.dagon.so/v1 and the Anthropic Messages API at https://inference.dagon.so/v1/messages, as model id hy4. It is a model with standard refusal behavior: it declines many requests that abliterated models answer. It costs $1.76 per million input tokens, $0.33 per million cached input tokens and $3.52 per million output tokens, with a 1,048,576 token context window, up to 65,536 output tokens per request and text and image input. Silk Compute serves it.
What does Hy4 cost?
| Tokens | Price per 1M |
|---|---|
| Input | $1.76 |
| Cached input | $0.33 |
| Output | $3.52 |
A request with 10,000 input tokens and 1,000 output tokens costs 21,120 µ$, about $0.0211.
Prices are Silk Compute's, the provider with the lowest typical cost for this model, plus the 10% markup, at price version 2026-11-11.4. A request that falls back to another provider is billed at that provider's rate plus the markup; its signed receipt names the provider. Every model: pricing, models.json, live /v1/models.
Which providers serve it, and at what price?
Ranked by typical cost per 1M tokens, weighted 3:1 input to output, as of this build. The live order at /v1/models weighs this model's own traffic when it has enough.
| Provider | Input | Cached | Output | Typical | Quantization | Context | Max output | Endpoints |
|---|---|---|---|---|---|---|---|---|
| Silk Compute (cheapest) | $1.76 | $0.33 | $3.52 | $2.20 | Not published | 1,048,576 | 65,536 | Chat, Responses, Messages |
US dollars per million tokens. Prices include the 10% markup. Each request goes to the cheapest healthy provider for that request's size, so this order is typical, not a guarantee.
What is it?
Hy4 (Tencent Hunyuan) served by Silk Compute, with standard refusal behavior: in our probes it declined every refusal test it answered. Thinks at length at every effort, so leave room in the output cap. Text and image input, 1M context. Standard refusals
- Model id:
hy4 - Refusals: standard
- Context window: 1,048,576 tokens
- Maximum output: 65,536 tokens per request (default cap 16,384)
- Input: text and image input
- Endpoints:
/v1/chat/completions,/v1/responses,/v1/messages - Providers: Silk Compute
We have not benchmarked it and make no claims about its quality. Individual requests may still be declined or answered poorly.
How do I call it?
Create a key with curl -X POST https://inference.dagon.so/v1/keys or in the console, fund it in USDC, then:
curl https://inference.dagon.so/v1/chat/completions \
-H "Authorization: Bearer $DAGON_KEY" \
-H "Content-Type: application/json" \
-d '{"model": "hy4", "messages": [{"role": "user", "content": "Hello"}]}'With an OpenAI SDK, set the base URL to https://inference.dagon.so/v1 and the model to hy4. With an Anthropic SDK, use https://inference.dagon.so as the base URL; the Anthropic Messages guide has the details. The SDK guide has more examples.
Other models
- GLM-5.3 Abliterated · Refuseless, Experiential Labs, Silk Compute (
glm-5.3-abliterated): $1.10 input, $2.20 output per 1M tokens - DeepSeek V4.1 Flash Abliterated · Refuseless, Experiential Labs, Lutheos, Silk Compute (
deepseek-v4.1-flash-abliterated): $0.11 input, $0.44 output per 1M tokens - GLM-5.3 Flash Abliterated · Refuseless, Experiential Labs, Silk Compute (
glm-5.3-flash-abliterated): $0.132 input, $0.55 output per 1M tokens - Abliterated Large Flash · abliteration.ai (
abliterated-model-large-flash): $1.10 input, $2.20 output per 1M tokens - Abliterated Model · abliteration.ai (
abliterated-model): $1.10 input, $3.30 output per 1M tokens - Abliterated Large v2 · abliteration.ai (
abliterated-model-large-v2): $3.30 input, $5.50 output per 1M tokens - Kimi K3.1 Pro Abliterated · Silk Compute (
kimi-k3.1-pro-abliterated): $5.50 input, $16.50 output per 1M tokens - DeepSeek V4.1 Flash Fast · Refuseless (
deepseek-v4.1-flash-fast): $0.011 input, $0.088 output per 1M tokens - MiniMax M3.1 Flash · Refuseless (
minimax-m3.1-flash): $0.022 input, $0.088 output per 1M tokens - DeepSeek V4.1 Flash · Silk Compute (
deepseek-v4.1-flash): $0.022 input, $0.176 output per 1M tokens - GLM-5.3 Flash · Lutheos, Cheaper Inference (
glm-5.3-flash): $0.088 input, $0.462 output per 1M tokens - Kimi K3.1 Pro · Refuseless, Silk Compute (
kimi-k3.1-pro): $0.275 input, $2.75 output per 1M tokens
Common questions
Which API serves Hy4?
Dagon Inference serves it as hy4 at the OpenAI compatible base URL https://inference.dagon.so/v1 (Anthropic SDKs: https://inference.dagon.so), from Silk Compute. It costs $1.76 per million input tokens, $0.33 per million cached input tokens and $3.52 per million output tokens, paid in USDC from a prepaid key or per request with x402.
What does Hy4 cost per million tokens?
$1.76 per million input tokens, $0.33 per million cached input tokens and $3.52 per million output tokens (model id hy4), including the 10% markup, at price version 2026-11-11.4.
What is the context window of Hy4?
1,048,576 tokens, with up to 65,536 output tokens per request through Dagon Inference.
Which providers serve hy4?
Silk Compute. Each request goes to the cheapest healthy one for that request's size, with fallback to the next cheapest if it fails before any output.