Developed by iApp, launched with AIEAT. Part of OpenThai (formerly OpenThaiGPT).
OpenThai-SystemOne-Vision v1.4 for stock llama.cpp: the open Thai +
English System One decision model (0.86B parameters) that answers typed questions about a state, with or without a
screenshot, in one forward pass and without generating text. llama-server serves it through its /v1/systemone endpoint.
| file | size (bytes) | sha256 |
|---|---|---|
OpenThai-SystemOne-Vision-Q8_0-decision.gguf |
1,320,471,072 | 5b022149fbeccb906dd3f3579386ed65f8a8cfcaf2fdc92fda34f917b83659a0 |
mmproj-OpenThai-SystemOne-Vision-F16.gguf |
204,986,656 | 272385cd4a72182d59cbf7031e0bfc3cefca49cadb7ca34d7ba32bf33ab1c817 |
The first file is the text tower at Q8_0 with the decision readout (one row per option label, taken from the model's slot head, plus the per-type temperatures). The second is the vision tower and merger at F16; it is only needed when a request carries an image.
Later checkpoints are added as new commits. Pin the commit you tested if your app must not change behind its users' backs.
Tested with llama.cpp build b12242.
llama-server -m OpenThai-SystemOne-Vision-Q8_0-decision.gguf \
--mmproj mmproj-OpenThai-SystemOne-Vision-F16.gguf -c 32768 --parallel 4 -ngl 99
curl -s http://localhost:8080/v1/systemone -H 'content-type: application/json' -d '{
"state": "os: windows\napp: CalculatorApp.exe\ngoal: กดปุ่มบวก",
"questions": {"target": {"type": "choice",
"instructions": "Which element should be acted on to achieve the goal?",
"criteria": {"0": "button \"Seven\"", "1": "button \"Plus\"", "2": "button \"Equals\"", "none": "none of these"}}}
}'
The answer holds the chosen option and a probability for every option. Question types: choice (up to 255 options),
noul (yes / no; only instructions) and score (criteria is a list of 2 to 10 level names). The prompt layout is inside
the GGUF, so the client sends the request as it is.
This build has no point head: llama-server does not expose what that head reads. A desktop app clicks by asking a choice
question over the controls it reads from the screen through the accessibility API, as in the example above:
os: windows | macos | linux, app: <program>, goal: <the command>, one per line;button "Save", menuitem "View", listitem "General"),
in reading order, with "none": "none of these" last;What to expect from v1.4, measured with the PyTorch weights on our own Windows 11 test screens (about 1,650 targets, lists of up to 120 controls, no screenshot); this build picked the same options in our checks:
| the command | right control chosen |
|---|---|
names the control as written on screen, in English or after a Thai verb (กดปุ่ม Save) |
99 % |
| a natural English sentence | 96 % |
| a natural Thai sentence, English labels on screen | 86 % |
On macOS control lists (560 targets, os: macos) v1.4 chose the control named in the command 99 % of the time. Linux uses
the same format and is not measured yet.
Known weakness: when the thing asked for is not on the screen, v1.4 seldom answers none (8 to 15 % of such cases) and
picks some control instead, often with a high probability. Keep a check in the app before it clicks, for example that the
chosen control's name is related to the command, and ask the user when it is not. The next revision is trained for this.
Apache-2.0, as the source model, built on Qwen3.5-0.8B-Base. Developed by iApp, launched with AIEAT.
iapp/OpenThai-SystemOne-Vision