32B competitive-coding fine-tune
According to the model card, OlympicCoder-32B is a 32B code model fine-tuned from Qwen2.5-Coder-32B-Instruct on decontaminated Codeforces data.
Open Source Model Profile · open-r1
OlympicCoder-32B
OlympicCoder-32B is a 32.76B-parameter Qwen2 code fine-tune from open-r1 for competitive programming on decontaminated Codeforces data.
- Publisher
- open-r1
- Task
- text-generation
- Model type
- qwen2
- License
- apache-2.0
- Library
- transformers
- Publication status
- Accepted · not indexed
Model overview
OlympicCoder-32B is published by open-r1 as a Qwen2 text-generation code model. The captured configuration identifies Qwen2ForCausalLM and Safetensors metadata reports 32,763,876,352 parameters. According to the model card, it was fine-tuned from Qwen2.5-Coder-32B-Instruct on decontaminated Codeforces data, and card data records apache-2.0.
Recorded capabilities
LiveCodeBench-oriented evaluation
The model card describes LiveCodeBench v4_v5 with 268 Python problems, while noting the model trained only on C++ solutions and is partly out-of-domain.
Long chain-of-thought setup
According to the model card, the chat template prefills a <think> token for long chain-of-thought output, with documented sampling guidance.
Use cases in the source record
- Competitive-coding experiments in the LiveCodeBench and Olympiad-style areas the model card names.
- Chain-of-thought code-generation trials using the publisher's documented Transformers pipeline and <think>-prefilled chat template.
Limitations and unknowns
- No evaluation results were extracted from this record.
- No context-window value was extracted from this record.
- According to the model card, Python benchmark behavior is partly out-of-domain because training used C++ solutions.
- Provider state is historical snapshot data and should be refreshed before being presented as current.
Source and provenance
Source: open-r1/OlympicCoder-32B
Captured: Unknown. Processed: 2026-09-07T19:35:27.712759+00:00.
Model Card for OlympicCoder-32B OlympicCoder-32B is a code model that achieves very strong performance on competitive coding benchmarks such as LiveCodeBench andthe 2024 International Olympiad in Informatics. Repository: https://github.com/huggingface/open-r1 Blog post: https://huggingface.co/blog/open-r1/update-3 Model description Model type: A 32B parameter model fine-tuned on a decontaminated version of the codeforces dataset. Language(s) (NLP): Primarily English License: apache-2.0 Finetuned from model: Qwen/Qwen2.5-Coder-32B-Instruct Evaluation We compare the performance of OlympicCoder models on two main benchmarks for competi…
F001F002F003F004F005F006F007F009F010F011F013F014F015F016F017F018F019F020