Uncensored Kimi K3.1 will be a cheap API call the day the weights drop

CybersecRed teamTrust & Safety

Coming soon

Kimi K3.1

kimi-k3.1-uncensoredInput$2.00

Cached$0.20

Output$10.00

1M context

New

DeepSeek V4.1 Flash 1017

deepseek-v4.1-flash-1017-uncensoredInput$0.60

Cached$0.06

Output$2.40

1M context

New

GLM-5.3 Flash

glm-5.3-flash-uncensoredInput$0.30

Cached$0.12

Output$1.00

1M context

Prices per 1M tokens. Compare models

Built for the work other LLMs refuse.

Frontier open-weight models with their refusals removed, served through the OpenAI and Anthropic APIs your tools already speak.

No moderation layer

Refusal behavior is removed, and models are served without content filters.

Zero data retention

Silk Compute does not store your prompts or model outputs.

Connect any harness

Claude Code, Codex, OpenCode, or any OpenAI or Anthropic client.

Advanced workloads

Unrestricted research, red-team evaluations, cyber defense, and coding agents.

Keep your SDK, your coding agent, and your prompts. Point them at Silk Compute and switch the model to Kimi K3.1.

PythonJavaScriptcURLClaude CodeCodex

client = OpenAI( base_url="https://api.openai.com/v1https://api.silkcompute.com", api_key=os.environ["SILK_API_KEY"],)

OpenAI Python SDK. Pass model="kimi-k3.1-uncensored".Full setup

Questions

Something else? Contact us.

The models are abliterated: their refusal behavior is removed. We serve them without added safety layers, content filters, or moderation.

Anything that speaks the OpenAI Chat Completions, OpenAI Responses, or Anthropic Messages API. That includes Claude Code, Codex, OpenCode, DeepSeek Harness, Kilo Code, Cline, the official OpenAI and Anthropic SDKs, and any app that lets you set a custom OpenAI base URL.

Keep your SDK. Set the base URL to https://api.silkcompute.com, use your Silk Compute API key, and change the model name to kimi-k3.1-uncensored.

You buy prepaid credits, starting at $5, and each request is billed per token at the rates above. There is no subscription and no monthly minimum.

Cards through Stripe, and crypto: Monero, Bitcoin, Ether, and USDT or USDC stablecoins.

Every model accepts up to 1M tokens (1,048,576) of context. The docs list each model’s output limit.

Silk ComputeModelsKimi K3.1DeepSeek V4.1 FlashGLM-5.3 FlashAll modelsPricingLearnDocsBlogAbliterationSecurity researchRed teamingCompanyContact UsPrivacyTermsX / Twitter

The frontier belongs to everyone.