X-Coder-SFT-376k: Fully Synthetic Competitive Programming Tasks and Solutions

887,321 synthetic records across 423,883 unique queries form this dataset for advancing competitive programming. It is designed for supervised fine-tuning and curated by state-of-the-art reasoning models. The dataset was authored by IIGroup and last updated on February 7, 2026.

Use Cases

• Supervised fine-tuning of code generation models based on synthetic tasks and solutions.
• Training foundational models for code reasoning using the described large-scale synthetic queries.
• Benchmarking model performance on competitive programming problems generated by reasoning models.

Strengths

• Large scale with 887,321 total synthetic records.
• Contains 423,883 unique queries, suggesting diversity.
• Designed specifically for supervised fine-tuning from a cold start.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count for the specific 'X Coder Sft 376K' subset is unknown, which may limit suitability assessment.
• Data is fully synthetic, which may not perfectly reflect real-world competitive programming distributions.

Provenance

Source

IIGroup via Hugging Face.

Collection Method

Curated by state-of-the-art reasoning models; described as fully synthetic.

Freshness

Last updated 2026-02-07 11:04:07; freshness should be verified.