TIME: Temporally Intelligent Meta-reasoning Engine for Context-Triggered Explicit Reasoning

TL;DR

TIME is a framework that enables language models to perform context-triggered explicit reasoning in dialogue, improving efficiency and responsiveness to temporal cues by learning a control policy for reasoning bursts.

Contribution

It introduces a novel behaviorally aligned approach that triggers explicit reasoning based on temporal context, reducing token usage and enhancing model responsiveness.

Findings

TIME improves reasoning benchmark scores across multiple scales.

Explicit reasoning tokens are reduced by roughly an order of magnitude.

Models become more compact and context-sensitive in their reasoning behavior.

Abstract

Reasoning-oriented language models typically expose explicit reasoning as a long, front-loaded chain of "thinking" tokens before the main output, either always enabled or externally toggled at inference time. Although this can help on arithmetic, coding, and other multi-step tasks, it is costly, weakens claim-level auditability, and does not allow the model to re-trigger explicit reasoning once presentation has begun. In dialogue, these limitations are compounded by weak sensitivity to temporal structure: unless time is explicitly stated in text, standard models treat replies separated by seconds and replies separated by weeks as equivalent. We introduce TIME (Temporally Intelligent Meta-reasoning Engine), a behavioral alignment framework that learns explicit reasoning as a context-triggered control policy rather than a fixed response mode. TIME augments dialogue with optional ISO 8601 <time> tags, tick events that represent silent time passage, and short <think> blocks that may appear anywhere in a response. Using a four-phase curriculum, including a small maximally diverse full-batch alignment stage, we train Qwen3 dense models to invoke brief, in-place reasoning bursts only when contextual cues warrant them, while keeping user-facing output compact. We also introduce TIMEBench, a diagnostic benchmark for evaluating reasoning from temporal cues in dialogue. Across 4B-32B scales, TIME improves TIMEBench scores over the corresponding base Qwen3 models in both thinking and no-thinking modes while reducing explicit reasoning tokens by roughly an order of magnitude. Beyond score improvements, TIME induces a distinct behavioral shift: explicit reasoning becomes more compact and more responsive to contextual cues. Code, training data, and benchmark artifacts are publicly available.