Ternary bonsai 2 is a 27b model that's 9x smaller than full precision and keeps 98.2% of benchmark…
ternary bonsai 2 is a 27b model that's 9x smaller than full precision and keeps 98.2% of benchmark performance. the frontier stopped being bigger and became smaller at 9x the efficiency.
the models that win production won't be the ones at the top of the leaderboard.
Context
PrismML's post of 17 September 2026 describes Ternary Bonsai 2 27B, based on Qwen3.8 27B, with ternary weights and FP16 group-wise scaling at 1.76 effective bits per weight, a 5.9GB footprint, 262K context, multimodal input and an Apache 2.0 license. PrismML claims it is more than 9x smaller than the full-precision baseline and keeps 98.2% of aggregate benchmark performance, an aggregate score of 83.9 across reasoning, math, coding, instruction following, vision and agentic tool use.
The 98.2% is PrismML's own aggregate retention against full-precision Qwen3.8 27B, vendor-reported and not independent, not a per-task result; per-benchmark figures are in a whitepaper that was not inspected. The frontier stopped being bigger and the leaderboard line are the author's opinion.
Related work
- Ternary-Bonsai-2-27B GGUF (Hugging Face) ↗Snippet only.
Watch next
- Independent evaluations on the same harness, and the whitepaper table.
Sources
- Bonsai 2 27B (PrismML, 17 Sep 2026)prismml.com
Provenance
The note above is reproduced unedited from the original post, first published on Threads on 18 September 2026 at 06:00 IST. Sources are the papers and datasets the note draws on.
View the original post ↗Embed this note
More notes
The air is now being asked to keep its own ledger
the air is now being asked to keep its own ledger: ecmwf’s aifs compo becomes the first ai model to forecast atmospheric composition globally every three hours, cleanair simulates 365 days of pm2.5 over china in ten seconds, and a unified framework maps six pollutants at one kilometer across the whole country. the air now files its own composition report.
read the note →The current is now being asked to draw its own map
the current is now being asked to draw its own map: china’s langya 2.0 predicts six ocean phenomena including internal waves and mesoscale eddies, a deep net called wenhai resolves eddies globally with air sea flux formulas built in, and scripps infers surface currents from the way temperature patterns deform in satellite images. the ocean now files its own circulation report.
read the note →The soil is now being asked to report its own carbon
the soil is now being asked to report its own carbon: a nix color sensor paired with generative data augmentation predicts soil organic carbon without a lab, random forest drives 74 percent of soil health mapping studies, and sentinel 2 tracks five year carbon change across france and italy from 922 samples. the dirt now files its own carbon account.
read the note →