37 min ago
••
Latest News
Latest news from Subscribe to LocalLLaMA's sources
1h ago
Granite 4.1: IBM’s 8B Model Is Competing With Models Four Times Its Size
1h ago
Qwen-27B as a Local Agent — It Actually Works Now
1h ago
I built a 5M model to see if it outperforms my 350M model...
2h ago
Can't replicate Reddit numbers with Qwen 27B on a 3090TI.
2h ago
Notes on what actually breaks when you run a coding agent on small local models
3h ago
Qwen-Scope: Official Sparse Autoencoders (SAEs) for Qwen 3.5 models
4h ago
How are you maintaining your AI apps post-launch? Model bugs vs engineering bugs, and what's your debugging stack?
5h ago
Tenstorrent TT-QuietBox 2 Specifications (Blackhole)
5h ago
A conversation about local LLMs with a senior government AI leader
5h ago
Qwen3.6-27B-Q6_K - images
7h ago
Where the goblins came from
7h ago
AMD Engineers directly seeking ROCm feedback
7h ago
DeepSeek V4 isn't beating Opus, but it doesn't need to
10h ago
Qwen3.6-27B 4.256bpw in full VRAM on a 5070 Ti with 50000 q4_0 context - not turbo!
11h ago
Larger Gemma-4/Qwen3.6
13h ago
You should probably disable algif kernel module this second if you run linux.
13h ago
I got a Qwen sticker lol
14h ago
"What do you guys even use local LLMs for?" Me: A lot
16h ago