chat with template is broken (wrong model, multiline doesn't work)

[hainkind]

Hi i tried using the chat with a template. despite other settings it's using gemini-2.5-flash. also exit or multiline does not work.

~/.config/io.datasette.llm 
❯ cat default_model.txt 
gemini/gemini-3-flash-preview⏎                                                                                                                                             
~/.config/io.datasette.llm 
❯ cat templates/test.yaml 
model: gemini/gemini-3-flash-preview
system: speak french
prompt: hi

~/.config/io.datasette.llm 
❯ llm chat -t test
Chatting with gemini/gemini-2.5-flash
Type 'exit' or 'quit' to exit
Type '!multi' to enter multiple lines, then '!end' to finish
Type '!edit' to open your default editor and modify the prompt
Type '!fragment $my_fragment [$another_fragment ...]' to insert one or more fragments
> hi
Salut !
> exit
D'accord. Au revoir !
> 
> quit
Au revoir !
> !exit
D'accord, à bientôt !
> !multi hi
whats up
!end






Aborted!

~/.config/io.datasette.llm took 1m25s 
❯ llm logs -c
# 2026-01-12T08:16:34    conversation: 01kermecb3yyv3thddk2z2qw5p id: 01kermees3f3j8sb383066ty03

Model: **gemini/gemini-2.5-flash** (resolved: **gemini-2.5-flash**)

## Prompt

hi
hi

## System

speak french

## Response

Salut !

# 2026-01-12T08:16:41

## Prompt

hi
exit

## Response

D'accord. Au revoir !

# 2026-01-12T08:16:48

## Prompt

hi
quit

## Response

Au revoir !

# 2026-01-12T08:16:52

## Prompt

hi
!exit

## Response

D'accord, à bientôt !



~/.config/io.datasette.llm 
❯ llm --version
llm, version 0.28

[hainkind]

ah ok figured out the issue with the model. i had LLM_MODEL set in my env. however i think the template value should take precedence or it would be nice to get info why that model was taken (default file/ env/ template etc)