This model was the CK-Pro-8B model mentioned in the paper Cognitive Kernel-Pro: A Framework for Deep Research Agents and Agent Foundation Models Training.

This model is finetuned using self-collected trajectories from the queries as presented in Table 2 of the paper.

Qwen-3-8B-CK-Pro achieves a Pass@1/3 score of 32.7%/38.2% on the full dev set of GAIA, and 40.3%/49.3% on the text-only subset of GAIA.

Downloads last month
143
Safetensors
Model size
8.19B params
Tensor type
F32
·
BF16
·
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support

Model tree for CognitiveKernel/Qwen3-8B-CK-Pro

Base model

Qwen/Qwen3-8B-Base
Finetuned
Qwen/Qwen3-8B
Finetuned
(211)
this model
Quantizations
1 model