Arc-Teacher-8B-f32-GGUF

Arc-Teacher-8B-f32-GGUF

Arc-Teacher-8B is a specialized 8B-parameter reasoning model based on Qwen3-8B, trained with Reinforcement Learning from Teacher (RLT) and Group Relative Policy Optimization (GRPO), excelling at generating detailed teaching traces for mathematical problem-solving (46% accuracy on AIME 2025) and demonstrating robust cross-domain transfer (69.2% accuracy on CRM-arena violation detection); it is fine-tuned from bespoke-stratos-17k data, capable of explaining its reasoning process step by step, facilitates both mathematical and cross-domain evaluation, and is released under the Apache 2.0 license for teaching, research, or deployment environments that can handle its substantial computational resource requirements.

Execute using Ollama

run ->

ollama run hf.co/prithivMLmods/Arc-Teacher-8B-f32-GGUF:BF16

Model Files

Quants Usage

(sorted by size, not necessarily quality. IQ-quants are often preferable over similar sized non-IQ quants)

Here is a handy graph by ikawrakow comparing some lower-quality quant types (lower is better):