AI Briefing#unsloth
The latest AI and developer news about #unsloth, with the original source and a short summary.
The latest AI and developer news about #unsloth, with the original source and a short summary.
Moving the monitor to the iGPU raised llama.cpp token speed from 50t/s to 55t/s.
On an RTX 5070 Ti 16GB and 32GB RAM, Qwen3.6-35B-A3B Q8_0 achieved 44 t/s.
Connecting Qwen3.6 to Claude Code on a dual 3090 setup cut the estimated 8-hour cost from $142 to under $4.
Unsloth has released a GGUF distribution of Kimi K2.6 along with related documentation.
On Gemma 4 26B-A4B GGUF, Unsloth quants top the charts by KLD in most cases.
Unsloth's Mistral Small 4 GGUF has been updated along with chat template fixes.
Uniform quantization breaks Qwen3.5's sensitive layers.
In the Qwen3.6-35B-A3B GGUF benchmark, Unsloth quants claim the best size efficiency relative to KLD.
Qwen3.6-35B-A3B-UD-Q2_K_XL handled 58 tool calls at a 98.3% success rate on a 16GB VRAM laptop.
GRaPE 2 was released in two models, 5B and 9B, emphasizing agentic reasoning.
NVIDIA has optimized Gemma 4 for local execution on RTX, DGX Spark, and Jetson.
Introduces how to efficiently fine-tune LLMs using coding agents by combining Unsloth with Hugging Face Jobs.
The Unsloth library can speed up LLM fine-tuning by up to 2x while reducing memory usage by 40%.