Gemma 4 12B
Fewer than 5 people recommend it so far, so there's no written verdict. Here is what commenters said.
3recommend
2warn against
Jun 29, 2026first mention
Sep 21, 2026latest mention
Why people recommend it
“On Gemma 4 12B, I am getting 220 ms per move”
runs on an old local PCon “Kev: Tiny Jev-like family of decision models built o…”
“Yesterday I did have success with Gemma-4-12b with 128k context.”
Show 1 more
“I use Gemma 4 12B in the 4-bit quantization for all sorts of vision tasks (image sorting, classification, description).”
Caveats and dissent
“seems more like a downgrade than an upgrade”
M1 MacBook Pro with 16GBswitched from Qwen2.5-14Bon “Qwen 3.6 27B is the sweet spot for local development”
“it just starts confusing itself really quickly when I give it coding tasks. And quickly starts failing tool calls.”
runs a 24GB M4 Pro MacBook Pro and a 3090on “Qwen 3.6 27B is the sweet spot for local development”
Often compared with
Who recommends it in their own words
Threads where it came up
Also written as Gemma 4 12b. Data: gemma-4-12b.json. Quotes are verbatim; the verdict and theme labels are written by a model from these comments, and every clause links to its sources.
All comments 6
“On Gemma 4 12B, I am getting 220 ms per move”
“I'm trying to use Gemma 4 12b for some workloads but I haven't mastered the model yet.”
“Yesterday I did have success with Gemma-4-12b with 128k context.”
“I use Gemma 4 12B in the 4-bit quantization for all sorts of vision tasks (image sorting, classification, description).”
“seems more like a downgrade than an upgrade”
“it just starts confusing itself really quickly when I give it coding tasks. And quickly starts failing tool calls.”