I ran the same todo-app prompt through the same agent harness a 3rd time.

Same model, same prompt. Only the process changed — and it changed everything. Run 2: 3h17m😲 (85 minutes of it burned by one subagent chasing a bug it introduced). Run 3: 43 minutes. 102 model calls,

Chris WAug 302

This is wild, never thought I'd be running a 27B-parameter model on my phone.

Running Alibaba_Qwen's Qwen 3.8 27B (1-bit) on my iPhone 17 Pro. gonna share the benchmarks very soon. Stay tuned! You can try running it on your own phone using the RunAnywhereAI apps available on

Sanchit mongaAug 3011

Is DeepSeek V4 Flash still the one to beat?

TonyD2Wild thinks so. Pound for pound he still has it over Qwen 3.8 Flash on the benchmarks. DeepSeek stays a big dog.

2WiLD ClipsAug 30

All made local with qwen 3.8 27b researching and generating lyrics and minimax music generating the song

SSHAug 302