Hy4 Xploded

hy4-xploded

Tencent's Hy4, uncensored: a million tokens of context, tool calling, reasoning and vision.

Modalities
Price per 1M
$3.20 / $6.40
input / output, $0.60 cached input
Context
1.05M
1,048,576 tokens

The uncensored Hy4 API

MagmaRouter serves an uncensored build of Hy4 under the model id hy4-xploded. It has 1.05M tokens of context, costs $3.20 per million input tokens and $6.40 per million output tokens, with cached input at $0.60, and supports tool calling, so it can drive an agent. One key works on the OpenAI-compatible endpoint (https://api.magmarouter.com/v1) and on the Anthropic-compatible one Claude Code uses, so the same Hy4 model plugs into Claude Code, Cline, opencode, SillyTavern, the OpenAI SDK and anything else that takes a custom base URL.

Hy4 uncensored against the rest of the line

ModelContextInputOutputIntelligenceToolsImages
GLM 5.3 Xploded1.05M$2.00$5.0045YesNo
GLM 5.3 Flash Xploded1.05M$0.30$1.0042YesYes
DeepSeek V4.1 Flash 1017 Xploded1.05M$1.25$2.5039YesYes
MiMo V2.6 Flash Xploded1.05M$1.00$2.0038YesYes
Qwen3.8 27B Xploded1.05M$1.50$5.0034YesYes
Qwen3.8 27B Ultrafast Xploded524k$1.75$5.5034YesYes
DeepSeek Flash Lite Xploded512k$0.36$1.4439YesYes
Hy4 Xploded1.05M$3.20$6.40-YesYes
Kimi K3.1 Pro Xploded1.05M$7.50$37.5055YesYes

Prices are per million tokens. Intelligence is the Artificial Analysis Intelligence Index of the base model. DeepSeek V4.1 Flash 1017 Xploded has no score of its own, so it shows DeepSeek V4.1 Flash's, the model it was tuned from. DeepSeek Flash Lite Xploded has no score of its own, so it shows DeepSeek V4.1 Flash's, the model it was tuned from.

What the uncensored tag actually means

hy4-xploded is tagged uncensored, which here means a loosely aligned finetune. It refuses far less than a frontier instruct model, but it is not abliterated: the refusal behaviour was reduced by training, not surgically removed, so it can still decline. Nothing about the tag promises it will answer anything.

Call it with curl

Shell
curl https://api.magmarouter.com/v1/chat/completions \
  -H "Authorization: Bearer sk-mr-..." \
  -H "Content-Type: application/json" \
  -d '{
    "model": "hy4-xploded",
    "messages": [{"role": "user", "content": "Hello"}]
  }'

Call it with the OpenAI SDK

Python
# pip install openai
from openai import OpenAI

client = OpenAI(
    base_url="https://api.magmarouter.com/v1",
    api_key="sk-mr-...",
)

resp = client.chat.completions.create(
    model="hy4-xploded",
    messages=[{"role": "user", "content": "Hello"}],
)
print(resp.choices[0].message.content)

In Claude Code

This model supports tool calling, so it can drive Claude Code. Set the context length to its real value of 1,048,576:

Shell
# Put these in your shell profile, then run: claude
export ANTHROPIC_BASE_URL=https://api.magmarouter.com
export ANTHROPIC_AUTH_TOKEN=sk-mr-...
export ANTHROPIC_MODEL=hy4-xploded
export ANTHROPIC_DEFAULT_HAIKU_MODEL=glm-5.3-xploded
export CLAUDE_CODE_MAX_CONTEXT_TOKENS=1048576

FAQ

What does the uncensored tag mean?

Uncensored here means a loosely aligned finetune that refuses much less than a frontier model, but can still decline. It is not the same as abliterated.

Does Hy4 Xploded store my prompts?

No. MagmaRouter stores the model id, token counts and cost, not the prompt or completion, and returns a signed receipt of the prompt hash.

How much does it cost?

$3.20 per million input tokens and $6.40 per million output tokens.

Is there an uncensored Hy4 API?

Yes. MagmaRouter serves Hy4 with the refusal training loosened (uncensored) as hy4-xploded, on an OpenAI-compatible endpoint at https://api.magmarouter.com/v1 and an Anthropic-compatible one for Claude Code.

What is the Hy4 context window?

1,048,576 tokens (1.05M).

Does Hy4 uncensored support tool calling?

Yes. It can call tools, so it can drive coding agents such as Claude Code, Cline and opencode.

Can Hy4 uncensored read images?

Yes. Send images as OpenAI image_url content parts, or paste them into the chat.

Can I use Hy4 uncensored in SillyTavern?

Yes. Add MagmaRouter as an OpenAI-compatible (Chat Completion) source with the base URL https://api.magmarouter.com/v1, your key, and hy4-xploded as the model.

The rest of the line