ZGCM-1: Fully Open 7B Foundation Model Excels in Math and Agentic Search with 4.2x Efficiency

ZGCM-1 is a fully open 7B foundation model trained from scratch with extreme data, system, and algorithmic efficiency, designed for math and agentic search with up to 256K context.

ZGCM-1 is competitive with frontier models orders of magnitude larger, such as Qwen3-235B-A22B and GLM-5.1, on challenging mathematical reasoning and agentic search suites.

Yes, it is fully open-source with weights, code, and data recipes, making it suitable for regional teams seeking cost-effective and resource-efficient solutions.

Key innovations include hybrid interleaved gated sliding-window and full attention, a stable FP8 Muon optimizer, progressive context scaling, and reformulation of interaction traces into Markov Decision Processes.