AdarshSingh7647/HETU-GLM-Z1-9B-MathReasoning-CotGen

TEXT GENERATIONPricing:Input $0.431 / Cached $0.0862 / Output $1.12Concurrent Unit Cost:1Model Size:9BQuant:FP8Context Size:32kTool Calling:SupportedPublished:Aug 26, 2026Architecture:Transformer Featherless Exclusive Cold

AdarshSingh7647/HETU-GLM-Z1-9B-MathReasoning-CotGen is a 9 billion parameter language model based on the GLM-Z1-9B-0414 architecture, specifically fine-tuned for advanced mathematical reasoning tasks. This model is distinguished by its Chain-of-Thought Generation (CotGen) method, where it is trained to produce a full reasoning process before generating its final answer. It excels in complex math problems across various benchmarks like AIME, GSM8K, and MATH-500, making it highly suitable for applications requiring detailed, step-by-step mathematical problem-solving.

Loading preview...

HETU-GLM-Z1-9B-MathReasoning-CotGen Overview

This model, developed by AdarshSingh7647 as part of the HETU (Hints Enable True Understanding) suite, is a 9 billion parameter language model built upon the zai-org/GLM-Z1-9B-0414 base architecture. Its primary focus is on advanced mathematical reasoning.

Key Capabilities

  • Mathematical Reasoning: Specifically trained and optimized for solving complex math problems.
  • Chain-of-Thought Generation (CotGen): Utilizes a unique training method where it generates a complete step-by-step reasoning process before providing the final output, enhancing transparency and accuracy.
  • Benchmark Performance: Evaluated and designed for strong performance on challenging math datasets including AIME, GSM8K, MATH-500, Omni-MATH, GPQA-Diamond, and MMLU.
  • Model Configuration: Provided as a merged model (base weights + LoRA adapter) in bf16 format, representing the final training checkpoint.

Good For

  • Complex Math Problem Solving: Ideal for applications requiring detailed solutions to mathematical challenges.
  • Educational Tools: Can be used to generate explanations and step-by-step solutions for math problems.
  • Research in AI Reasoning: A valuable tool for exploring and developing advanced reasoning capabilities in large language models, particularly in the mathematical domain.

Further details on its training setup, evaluation methodology, and comprehensive results can be found in the associated HETU paper.

Finetunes
1 models