It's insane how lobotomized Opus 4.6 is right now. Even Gemma 4 31B UD IQ3 XXS b

<table> <tr><td> <a href="https://www.reddit.com/r/singularity/comments/1tdft9u/poetiq_recursive_selfimprovement_delivers_new/"> <img alt="Poetiq: Recursive Self-Improvement

<!-- SC_OFF --><div class="md"><p>Looks like &quot;if old model get dumb, new model more smart!&quot; is actually what the strat is at Anthropic. </p> <p>If you spent a mint on

<!-- SC_OFF --><div class="md"><p>Benchmark: 58.4 vs 56.7 (beats GPT-5.4)</p> <p>License: Fully open (Apache 2.0) </p> <p>What it actually does: Runs 8-hour fully autonomous agent

<!-- SC_OFF --><div class="md"><p><a href="https://huggingface.co/unsloth/gemma-4-E2B-it-GGUF">https://huggingface.co/unsloth/gemma-4-E2B-it-GGUF</a></p> <p><a

<table> <tr><td> <a href="https://www.reddit.com/r/LocalLLaMA/comments/1srhtvg/where_we_are_in_a_year_everything_has_changed/"> <img alt="Where we are. In a year, everything has

<!-- SC_OFF --><div class="md"><p>I can't believe it, but I'm able to do my daily software development work on this model. We have a 500-700k line of code enterprise software

<!-- SC_OFF --><div class="md"><p>I run Gemma 4 26B-A4B locally via Ollama as part of a custom self-hosted AI platform. The platform stores every model interaction in SQLite,

<!-- SC_OFF --><div class="md"><p>Few days ago I posted about my <a href="https://www.reddit.com/r/LocalLLaMA/comments/1t94ito/i_have_deepseek_v4_pro_at_home/">DeepSeek V4 Pro</a>