llama.cpp
`server`: add `--reasoning-budget 0` to disable thinking (incl. qwen3 w/ enable_thinking:false)
#13771
Merged
Merged
`server`: add `--reasoning-budget 0` to disable thinking (incl. qwen3 w/ enable_thinking:false) #13771
ochafik
merged
13 commits into
ggml-org:master
from
ochafik:enable-thinking
server: fix/test add_generation_prompt
8a25f796
tools: enable hermes2/qwen chat logic even w/o tools
43b5626e
server: add --reasoning-format=disabled to disable thinking (incl. qw…
b457f89e
Update README.md
df25e6bb
Add models/templates/Qwen-Qwen3-0.6B.jinja
b6eb0a5c
update --reasoning-format={disabled -> nothink} as suggested
cdea6a9b
fix command r7b's nothink w/ official template
473c01e5
ochafik
changed the title
rewrite docs as list as suggested
6b9efe7b
Update common/chat.cpp
355b38c9
Merge branch 'enable-thinking' of github.com:ochafik/llama.cpp into e…
fe6022f4
const char* return for chat enum name helpers
8547fccf
ngxson
commented
on 2025-05-25
switch to --reasoning-budget flag
9cdeebe8
Merge branch 'fix-gen-prompt' into enable-thinking
91623806
ngxson
approved these changes
on 2025-05-25
ngxson
changed the title
ngxson
changed the title 
Reviewers
Assignees
No one assigned
Labels
testing
examples
python
server
Milestone
No milestone
Login to write a write a comment.
Login via GitHub