cyberneurova

You can run the quantized version of Qwen3.8; the model size is 10GB, and it outperforms most models in day-to-day tasks as well as advanced tasks. https://x.com/cyberneurova/status/2088713510469386518?s=46

Desktop app have been updated too https://cyberneurova.ai/downloads

Cyberneurova cli now supports Qwen 3.8 npm install -g cyberneurova-cli

Qwen 3.8 2t model is now available for paid users starting from premium - pro max , I’m working on the 27 b model that will serve from free users to also pro max https://cyberneurova.ai

cyberneurova-Qwen3.8-27B — GGUF An uncensored, neutral 27B assistant. It answers directly, with no disclaimers or caveats — the tone is entirely yours, set by your own system prompt. - 262K context, reasoning, tool-calling, multilingual - Sees images (just add mmproj-F16.gguf) - 92% pass@1 on an executable coding benchmark - Runs locally on llama.cpp, Ollama, LM Studio, or Jan from q2_k 10gb in size to q8 20gb Pick a quant — Q4_K_M (~17 GB) is the sweet spot, up to Q8_0 for max quality: https://huggingface.co/cyberneurova/cyberneurova-Qwen3.8-27B-GGUF

You can refer your friends to earn tokens https://cyberneurova.ai/referrals

For credibility, and as we wait for the Android app to be approved, we have open-sourced our mobile application. This comes with the introduction of agents running directly on your mobile phone without relying on any third-party servers. Check the repo and feel free to contribute to make the app better. https://github.com/CyberNeurova/cyberneurova-mobile-oss

For the mean time IOS app has been available on App Store for a week now https://apps.apple.com/ke/app/cyberneurova/id6776460478

what you should expect ✈️

The Android app restructure is complete; the agentic work is fully done on your Android device. The app has been published and is waiting for review 🍌 https://cyberneurova.ai

Cyberneurova through hunters just received its first payout from bug bounties 🙂

Android app should be out soon, and we believe it will be better. For Android, we will be supporting agentic tasks, and it will come with terminal support. This will allow anyone with just an Android device to be able to run commands and do crazy tasks like building apps, security testing, etc. https://cyberneurova.ai

deepseek weights are also available on huggingface https://huggingface.co/cyberneurova/cyberneurova-Deepseek-v4-flash

New model live: cyberneurova DeepSeek v4 Flash Our new reasoning + code flagship just landed on CyberNeurova. Deep coding, refactors & tool use Full reasoning — thinks before it answers Massive context for whole codebases & long docs Direct answers, no lectures Available now on Premium, Pro & Pro Max. the model has been fully abliterated 👉 https://cyberneurova.ai

We are working to make GLM 5.2 accessible to free users on mobile. At the same time, Kimi 3 will be available to Pro users by the end of this week; currently, Kimi 3 is accessible only to Pro Max users.

Check all models and services are back up https://cyberneurova.ai

Payments have been affected too; for subscriptions, only Apple payments are available at the moment.