Paper 2603.00889
CHIMERA: Compact Synthetic Data for Generalizable LLM Reasoning
- Published
- Mar 2026
- Research lab
- Independent
- Citations
- 2
- GitHub
- Not linked
01 In brief
Summary
The paper introduces CHIMERA, a compact synthetic dataset of 9,225 samples designed to address data-centric challenges in LLM reasoning post-training: cold-start data scarcity, limited domain coverage, and the annotation bottleneck.
The dataset is built via a three-stage pipeline: subject expansion using GPT-5 to create a hierarchical taxonomy of 1,179 topics across 8 scientific disciplines, problem generation with GPT-5, and solution synthesis using Qwen3-235B-A22B-Thinking-2507 to produce long Chain-of-Thought trajectories.
Quality is ensured through automated cross-validation by GPT-5 and o4-mini.
Post-training a Qwen3-4B model with SFT and RL on CHIMERA yields strong performance on benchmarks like GPQA-Diamond, AIME 24/25/26, HMMT 25, and HLE, approaching or matching much larger models such as DeepSeek-R1 and Qwen3-235B.
Analysis shows CHIMERA problems are more challenging than existing synthetic datasets (37.5% base accuracy vs.
76-88% for others), and LLM-generated problems are rated higher in quality than human-curated ones in blind evaluations.
Contamination analysis confirms negligible lexical overlap with test benchmarks.
02 From the paper
Abstract
Large Language Models (LLMs) have recently exhibited remarkable reasoning capabilities, largely enabled by supervised fine-tuning (SFT)- and reinforcement learning (RL)-based post-training on high-quality reasoning data. However, reproducing and extending these capabilities in open and scalable settings is hindered by three fundamental data-centric challenges: (1) the cold-start problem, arising from the lack of seed datasets with detailed, long Chain-of-Thought (CoT) trajectories needed to initialize reasoning policies; (2) limited domain coverage, as most existing open-source reasoning datasets are concentrated in mathematics, with limited coverage of broader scientific disciplines; and (3) the annotation bottleneck, where the difficulty of frontier-level reasoning tasks makes reliable human annotation prohibitively expensive or infeasible. To address these challenges, we introduce CHIMERA, a compact synthetic reasoning dataset comprising 9K samples for generalizable cross-domain reasoning. CHIMERA is constructed with three key properties: (1) it provides rich, long CoT reasoning trajectories synthesized by state-of-the-art reasoning models; (2) it has broad and structured coverage, spanning 8 major scientific disciplines and over 1K fine-grained topics organized via a model-generated hierarchical taxonomy; and (3) it employs a fully automated, scalable evaluation pipeline that uses strong reasoning models to cross-validate both problem validity and answer correctness. We use CHIMERA to post-train a 4B Qwen3 model. Despite the dataset's modest size, the resulting model achieves strong performance on a suite of challenging reasoning benchmarks, including GPQA-Diamond, AIME 24/25/26, HMMT 25, and Humanity's Last Exam, approaching or matching the reasoning performance of substantially larger models such as DeepSeek-R1 and Qwen3-235B.