merge

Name: Qwen2.5-7B-Sky-R1-Mini
Availability: InStock

This is a merge of pre-trained language models created using mergekit.

Merge Details

Merge Method

This model was merged using the SCE merge method using Qwen/Qwen2.5-7B-Instruct-1M as a base.

Models Merged

The following models were included in the merge:

open-thoughts/OpenThinker-7B
deepseek-ai/DeepSeek-R1-Distill-Qwen-7B
NovaSky-AI/Sky-T1-mini
open-r1/OpenR1-Qwen-7B

Configuration

The following YAML configuration was used to produce this model:

yaml

models:
  # Pivot model
  - model: Qwen/Qwen2.5-7B-Instruct-1M
  # Target models
  - model: open-r1/OpenR1-Qwen-7B
  - model: open-thoughts/OpenThinker-7B
  - model: NovaSky-AI/Sky-T1-mini
  - model: deepseek-ai/DeepSeek-R1-Distill-Qwen-7B
merge_method: sce
base_model: Qwen/Qwen2.5-7B-Instruct-1M
parameters:
  select_topk: 1.0
dtype: bfloat16

Open LLM Leaderboard Evaluation Results

Detailed results can be found here

Metric	Value
Avg.	7.37
IFEval (0-Shot)	23.05
BBH (3-Shot)	8.90
MATH Lvl 5 (4-Shot)	2.95
GPQA (0-shot)	5.26
MuSR (0-shot)	1.27
MMLU-PRO (5-shot)	2.81

bunnycore/Qwen2.5-7B-Sky-R1-Mini