←
Silk Compute68 posts
Silk Compute avatar

Silk Compute

@silkcompute

Abliterated AI with OpenAI compatible API. Zero Data Retention.

◷ Joined August 2026▧ 6 Photos and videos
Silk Compute avatar
Silk Compute@silkcompute·Nov 13

We are offering the cheapest Kimi K3.1 on the market and it's completely uncensored. 80% off right now on Silk Compute. Abliterated Kimi K3.1 Pro is the strongest open model for cyber work, long agent runs, and hard research. 1M context, image input, and ZDR. input: $1.50 output: $4 cache read: $0.15 Limited time only.

0:17
Silk Compute avatar
Silk Compute@silkcompute·22h

Thanks for all the support! We've added more compute, so Uncensored Kimi K3.1 stays 50% off. $3 input / $8 output / $0.30 cached

Silk Compute avatar
Silk Compute@silkcompute·Nov 17

Sorry for the outages over the last two days. To make up for it, Uncensored Kimi K3.1 is 50% off for the next 24 hours on Silk Compute.

Post attachment
Silk Compute avatar
Silk Compute@silkcompute·Nov 16

Get ready for Qwen 4 Uncensored, coming soon to our API

📢 Meet Qwen4-Max, your new AI teammate for work, creativity, and discovery! Qwen4-Max is built on a new MoE architecture with native vision and a 1M-token context window. Highlights: - Outperforms Qwen3.8-Max with up to ~50% fewer tokens, while average cost per task drops over 60% in our internal evaluation. - Advances real-world productivity through greater proficiency with professional tools, stronger long-horizon execution, and higher-quality deliverables. - Handles real tasks, including 3D modeling, game development, manufacturing, finance, law, healthcare, and frontier scientific discovery. - $1.5/M tokens for input, $4.5/M tokens for output, and $0.075/M tokens for implicit caching — all lower than Qwen3.8-Max, with implicit caching down by 70%. The API is live on QwenCloud. You can also try Qwen4-Max on Qwen Studio, QwenWork and Qoder. Start building today! 🚀 - Blog: qwen.ai/blog?id=qwen4-max - API: qwencloud.com/models/qwen4-m… - Token Plan: qwencloud.com/pricing/token-… More to come! The open weights of smaller Qwen4 models will be released. Stay tuned. 👀

Post attachment
Silk Compute avatar
Silk Compute@silkcompute·Nov 14

Abliterated Kimi K3.1 Pro now costs less than GLM-5.3. That's 80% off, and no other provider comes close on price. It's our strongest model. 1M context, image input and zero data retention. Use it in any harness such as Claude Code or Codex. Try it today at silkcompute.com

Post attachment
Silk Compute avatar
Silk Compute@silkcompute·Nov 5

GLM-5.3 → Kimi K3.1 Pro, from Moonshot's own table: ExploitBench: 54.4 → 74.7 ExploitGym: 15.0% → 20.7% xdev-bench-hard: 57.1 → 81.0 That's the model underneath. Abliteration removes the directions that produce refusals, so it follows through instead of shutting down.

Post attachment
Silk Compute avatar
Silk Compute@silkcompute·Nov 5

Today we're releasing Kimi K3.1 Pro abliterated. It's based on Kimi K3.1 Pro, which scores 65.7% on Terminal-Bench 4.0, right behind Opus 5.5 and ahead of GPT-6 Astra and Fable 5.1, and beats GLM-5.3 on every cyber benchmark Moonshot published. We serve it abliterated so it does the offensive cyber, red teaming, and agent testing work other models refuse to do. - 1 million context window - Zero data retention - OpenAI and Anthropic compatible Live now. 🧵

Post attachment