Kiwen1.2-27B
Kiwen1.2 is a fine-tuned version of Qwen3.8-27B, trained on long chain-of-thought reasoning traces from Kimi K3 — with a particular focus on reasoning, coding, tool use, and instruction following. But the goal wasn't simply to make the model think longer. The goal was to make it think better.
This model isn't an attempt to build the biggest model.
It's an attempt to make a 27-billion-parameter model think harder, follow instructions better, and act more reliably.
News
RL in internal workflow by RLVR
Internal benchmark
| Rank | Model | Pass Rate | Partial Rate |
|---|---|---|---|
| 1 | Kiwen1.2 | 81.2% | 89.62% |
| 2 | gemini-3.7-flash | 81.0% | 89.8% |
| 3 | gemini-3.6-flash | 80.0% | 88.9% |
| 4 | Kiwen1.1 | 78.1% | 87.85% |
the benchmark includes 100 internal tasks for agent workflow
Usage
from transformers import AutoModelForCausalLM, AutoTokenizer
m = "beyoru/Kiwen1.1-27B"
tok = AutoTokenizer.from_pretrained(m)
model = AutoModelForCausalLM.from_pretrained(m, dtype="auto", device_map="auto")
msgs = [{"role": "user", "content": "Natalia sold clips to 48 friends in April, "
"and half as many in May. How many total?"}]
ids = tok.apply_chat_template(msgs, add_generation_prompt=True,
enable_thinking=True, return_tensors="pt").to(model.device)
print(tok.decode(model.generate(ids, max_new_tokens=4096)[0][ids.shape[-1]:]))
Set enable_thinking=False for extraction, classification and formatting tasks.
The model was trained with both modes and respects the flag.
Serving with SGLang:
python -m sglang.launch_server --model-path beyoru/Kiwen1.2-27B \
--context-length 262144
Notes:
- Found some issues in response quality when using with dflash2 - make sure you don't use this model with it
Citation
@misc{kiwen27bk3,
title = {Kiwen1.2-27B},
author = {beyoru},
year = {2026},
url = {https://huggingface.co/beyoru/Kiwen1.2-27B}
}
License & Attribution
Built on:
- Qwen3.8-27B — Apache-2.0
Kiwen1.2-27B is released under the Apache-2.0 license.