← all releases
Nous Research·Jan 6, 2026·v1.0·open source
NousCoder-14B
NousCoder-14B is a competitive programming model developed by Nous Research, fine-tuned from Qwen3-14B using reinforcement learning. It achieved a Pass@1 accuracy of 67.87% on LiveCodeBench v6, surpassing the baseline accuracy of 60.79% from Qwen3-14B. The model was trained on 24,000 verifiable coding problems over four days using 48 B200 GPUs.
MLLLMsTestingCoding
overview
NousCoder-14B is a competitive programming model developed by Nous Research, fine-tuned from Qwen3-14B using reinforcement learning. It achieved a Pass@1 accuracy of 67.87% on LiveCodeBench v6, surpassing the baseline accuracy of 60.79% from Qwen3-14B. The model was trained on 24,000 verifiable coding problems over four days using 48 B200 GPUs.
key features
- 01Competitive programming specialization
- 02Post-trained on Qwen3-14B via reinforcement learning
- 03Achieved 67.87% Pass@1 accuracy on LiveCodeBench v6
- 04Trained on 24,000 verifiable coding problems
- 05Utilized 48 B200 GPUs for training
related productsbrowse all →
Step-Audio-R1open sourceStep-Audio-R1 is an advanced audio language model developed by StepFun AI, designed to enhance audio reasoning capabilities by grounding its reasoning in acoustic features. It introduces Modality-Grounded Reasoning Distillation (MGRD), an iterative training framework that shifts the model's reasoning from textual abstractions to acoustic properties, effectively addressing the 'inverted scaling' problem where performance degrades with longer reasoning. This model has demonstrated superior performance across various audio understanding and reasoning benchmarks, surpassing models like Gemini 2.5 Pro and achieving results comparable to Gemini 3 Pro.AIBuildAIopen sourceAIBuildAI is an AI agent that autonomously constructs AI models. Given a specific task, it initiates an agent loop to analyze the problem, design models, and execute training processes, all without human intervention.SINQopen sourceSINQ (Sinkhorn-Normalized Quantization) is a novel, fast, and high-quality quantization method designed to make any Large Language Model (LLM) smaller while preserving accuracy. It offers a plug-and-play, model-agnostic technique that delivers state-of-the-art performance for LLMs without sacrificing accuracy.