Kwaipilot/KAT-Coder-V2.5-Dev
Text Generation • 35B • Updated • 40.3k • 646
AI & ML interests
LLM, AI4SE, Agentic Coding
Recent Activity
View all activity
最好的35B模型，不输于Qwen3.8-27B
❤️ 1
1
#31 opened 14 days ago
by 
jgbrblmd
Presumably incorrect values from the QWEN original model
1
#30 opened 23 days ago
by 
kalle07
Coverage of Programming Languages in SFT and RL Training
1
#24 opened about 2 months ago
by 
enricos
Is AutoBuilder Open Source?
1
#23 opened about 2 months ago
by 
Behrens
As a non SWE agentic hobbyist I appreciate this model
❤️ 5
2
#21 opened about 2 months ago
by 
GreenMachine87
Does this preserve the scientific and reasoning mind of Qwen3.6 models or does that get scarificed for agentic coding
➕ 1
2
#22 opened about 2 months ago
by 
mayankiit04
Do the numbers differ from Ornith and Qwen graphs?
➕ 1
11
#1 opened about 2 months ago
by 
hgeist
No difference with base model (for me)...
5
#13 opened about 2 months ago
by 
jbourny
did you train it for 90 rl steps
1
#16 opened about 2 months ago
by 
inikishev
Wrong Qwen3.6-35B-A3B Benchmark
👍 1
2
#17 opened about 2 months ago
by 
XCurOS-35H
Community benchmark results discussion
👍 4
10
#12 opened about 2 months ago
by 
space-spike
Add community evaluation results for SWE-BENCH_VERIFIED, SWE-BENCH_PRO
4
#11 opened about 2 months ago
by 
nielsr
Does nobody use dark mode?
➕ 4
2
#10 opened about 2 months ago
by 
ercangorgulu
Ricardo-H
updated a
model about 2 months ago
My experience using and evaluating this model in coding tasks (the results are simply amazing).
👀❤️ 3
3
#6 opened about 2 months ago
by 
Am0gusIsSus
Parser error with claude code w/ or w/o shared template
3
#5 opened about 2 months ago
by 
xantrk
Parser error with claude code w/ or w/o shared template
3
#5 opened about 2 months ago
by xantrk
MTP support? Why no Vision?
20
#3 opened about 2 months ago
by 
unoid
Please specify attention mechanism (how KV cache is treated)
1
#2 opened about 2 months ago
by 
Reverger
Plans for FP8?
👍 3
1
#4 opened about 2 months ago
by 
SRZhu26