# TimeThink: Eliciting Compositional Reasoning in Timeseries Large Language Models

arXiv: https://arxiv.org/abs/2609.13457 · HTML: https://mlsanity.com/paper/2609.13457/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Placed in Time series reasoning, one of 17 directions in LLM reasoning.
- Builds on 17 of its 17 references that are in the map.
- First posted on arXiv 11 Sep 2026.

## Placed in

- LLM reasoning / [Time series reasoning](https://mlsanity.com/lr/time-series-reasoning/)
- Multimodal LLMs / vision-language models / [Chart understanding and anomaly detection](https://mlsanity.com/mllm/chart-understanding-and-anomaly-detection/)

## Builds on (most cited first, 17 of 17)

- [DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models](https://mlsanity.com/paper/2402.03300/) (2024)
- [HybridFlow: A Flexible and Efficient RLHF Framework](https://mlsanity.com/paper/2409.19256/) (2024)
- [WizardLM: Empowering large pre-trained language models to follow complex instructions](https://mlsanity.com/paper/2304.12244/) (2023)
- [Large Language Models Are Zero-Shot Time Series Forecasters](https://mlsanity.com/paper/2310.07820/) (2023)
- [ChatTS: Aligning Time Series with LLMs via Synthetic Data for Enhanced Understanding and Reasoning](https://mlsanity.com/paper/2412.03104/) (2024)
- [Language Models Still Struggle to Zero-shot Reason about Time Series](https://mlsanity.com/paper/2404.11757/) (2024)
- [Time-MQA: Time Series Multi-Task Question Answering with Context Enhancement](https://mlsanity.com/paper/2503.01875/) (2025)
- [TimeSeriesExam: A time series understanding exam](https://mlsanity.com/paper/2410.14752/) (2024)
- [SensorLM: Learning the Language of Wearable Sensors](https://mlsanity.com/paper/2506.09108/) (2025)
- [Time-R1: Towards Comprehensive Temporal Reasoning in LLMs](https://mlsanity.com/paper/2505.13508/) (2025)
- [By My Eyes: Grounding Multimodal Large Language Models with Sensor Data via Visual Prompting](https://mlsanity.com/paper/2407.10385/) (2024)
- [HEARTS: Benchmarking LLM Reasoning on Health Time Series](https://mlsanity.com/paper/2603.06638/) (2026)
- [Time2Lang: Bridging Time-Series Foundation Models and Large Language Models for Health Sensing Beyond Prompting](https://mlsanity.com/paper/2502.07608/) (2025)
- [STReasoner: Empowering LLMs for Spatio-Temporal Reasoning in Time Series via Spatial-Aware Reinforcement Learning](https://mlsanity.com/paper/2601.03248/) (2026)
- [LENS: LLM-Enabled Narrative Synthesis for Mental Health by Aligning Multimodal Sensing with Language Models](https://mlsanity.com/paper/2512.23025/) (2025)
- [LLaTiSA: Towards Difficulty-Stratified Time Series Reasoning from Visual Perception to Semantics](https://mlsanity.com/paper/2604.17295/) (2026)
- [SenTSR-Bench: Thinking with Injected Knowledge for Time-Series Reasoning](https://mlsanity.com/paper/2602.19455/) (2026)

## Abstract (arXiv)

Timeseries multimodal large language models (TS-MLLMs) have recently begun leveraging the reasoning capabilities of large language models (LLMs) for question-answering tasks. However, these models often fail to capture dynamic temporal patterns, providing only implicit reasoning that lacks the underlying explanations critical for high-stakes applications like healthcare. While reinforcement learning (RL)-based timeseries language models aim to address this, they often fall short because they are trained on narrow, in-distribution data and struggle with out-of-distribution compositional questions. To address these challenges, we present TimeThink, a synthetic framework for eliciting compositional timeseries reasoning. Core timeseries primitives (e.g., trend, seasonality) are domain-independent and can be deterministically generated. Guided by this premise, TimeThink first designs a synthetic data generator that produces atomic and composite question-answer pairs, providing objective ground truth with reasoning traces. Building on this framework, TimeThink employs a reinforcement learning with verifiable rewards (RLVR) training strategy that encourages explicit reasoning. Unlike template-reliant methods, this approach enables the model to learn the underlying logic of composition rather than simply imitating traces. Extensive experiments show that TimeThink, trained only on synthetic data, significantly outperforms strong baselines on both synthetic and real-world benchmarks.