1 min read

Free Supernova Coding Model Debuts as xAI Secures $10B, Grok 4 Fast Rivals Gemini 2.5 Pro

AI Models & Benchmarks

xAI Releases Grok 4 Fast with Performance Benchmarks
xAI unveiled Grok 4 Fast, showcasing detailed benchmarks that highlight significant improvements in speed and efficiency. The model is positioned as a top performer, rivaling Gemini 2.5 Pro in industry benchmarks, with transparent metrics for developer adoption.


Google Research Improves LLM Accuracy by Leveraging All Layers
Google’s new study demonstrates that utilizing all layers of large language models (LLMs) enhances accuracy and reliability. The findings suggest optimization techniques to maximize performance without architectural changes.


KaniTTS: Lightweight, High-Fidelity Text-to-Speech Model
NineNineSix.ai introduced KaniTTS, a 450M-parameter TTS model delivering real-time, low-latency speech synthesis with high fidelity. The model is optimized for conversational AI and accessibility applications.


Weekly AI Model Roundup (Sept 19)
A curated list of new/updated models from the past week, including:

  • Qwen3-Next-80B-A3B-Instruct (Alibaba) – Strong reasoning capabilities.
  • Magistral-Small-2509 (Mistral) – Compact, efficient model.
  • Ling-flash 2.0 & Ling-mini 2.0 (InclusionAI) – Multilingual focus.
  • MobileLLM-R1-950M (Meta) – Optimized for edge devices.
  • LocalAI v3.5.0 – Unified local AI framework with major updates.
  • Full list of releases

Coding & Developer Tools

Free Coding Model: Code-Supernova (Stealth Release)
Code-Supernova, a new free coding model, launched in alpha with:


Roo Code 3.28.4 Introduces Free Supernova Model
The latest Roo Code update (v3.28.4) includes:


Industry & Funding

xAI Raises $10B at $200B Valuation
Elon Musk’s xAI secured a $10 billion funding round, valuing the company at $200 billion. The investment reflects aggressive growth plans and heightened competition in the AI sector.