A pilot at Biome Makers with the real data of two clients of its soil-analysis platform. An open 122-billion-parameter model on two GPUs. CA3 keeps one memory per client and region; account questions go to a program that computes every figure from the client's own tables, with the memory switched off.
Results recorded on 15 and 17 November 2026
What a grower asks on a call is written in no document: a mean across lots, a count of fields, whether a gap is real or chance. These questions go to an account program, with the client memory switched off: it computes each figure on the client's tables, and the model writes the answer with the result in front of it. When the data cannot answer, it says so.

Every lot carries a soil map, pixel by pixel: here the best and the weakest lot of one farm in phosphorus, 70 and 35 on average. A question across all the lots is written nowhere: it is computed over the whole set. Real client maps, shown without names and without coordinates.
CA3 learns each client's own sheets (production, varieties, treatments) on top of what it already knows. Each region or farm has its own memory, and each question goes to the one it is about.
Learning run of 15 November 2026: two regions and two farms of the two clients, questions no one had seen. With CA3 answering what it learned and document search the rest, 27 to 36% less text went into the model.
The memory of a whole region, in one client, and of a whole farm, in the other, was removed on request. Asked for its figures afterwards, the model gave none of them. The client's other memory stayed as it was.
Before CA3, the pilot asked whether an open model could write the AI analysis of the platform's farm reports. Qwen3.8-27B, on one GPU, wrote the analysis of seven farms with no wrong figure among the 163 checked; the commercial model it was compared with had one in 112. CA3 did not run in this phase.
Two clients' real data, two runs: one for what CA3 learns and removes, one for the answers.
The account program reads the client's tables through a fixed catalogue of computations; a figure in the answer that is not in the result is replaced by “not available” before it leaves. The whole run, six ways of answering and 810 answers, used 19.72 USD of GPU time: about 2.4 US cents per answer.
| Measured | Asked | Account program | Document search | Model alone, no data |
|---|---|---|---|---|
| Account questions, right | 76 | 75 | 15 | 3 |
| Statistical questions, right | 12 | 11 | 0 | 1 |
| Declined where the data cannot answer | 14 | 14 | 11 | 8 |
| Answers with a wrong figure | 130 | 0 | 72 | 111 |
| Input tokens per question, mean | – | 5,614 | 7,090 | – |
| Seconds per question, mean | – | 8.74 | 8.44 | – |
A pilot on your own platform: your results, maps and sheets, one private memory per client, and four things measured on your data: right answers, invented figures, refusals where the data cannot answer, and cost per answer.
ContactAll use casesResults from recorded, pre-registered tests. The mechanism is not shown. Model names are for identification only. Not produced or endorsed by Alibaba or NVIDIA. CA3 is developed by Alberto Acedo, PhD, at Biome Makers Inc.