# Agogic: Performance-Timed Music Tokens for LLM-Native Text-to-Symbolic-Music Generation

arXiv: https://arxiv.org/abs/2608.03999 · HTML: https://mlsanity.com/paper/2608.03999/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 33 of its 116 references that are in the map.
- First posted on arXiv 4 Aug 2026.

## Builds on (most cited first, 30 of 33)

- [Qwen3 Technical Report](https://mlsanity.com/paper/2505.09388/) (2025)
- [Qwen-Audio: Advancing Universal Audio Understanding via Unified Large-Scale Audio-Language Models](https://mlsanity.com/paper/2311.07919/) (2023)
- [Jukebox: A Generative Model for Music](https://mlsanity.com/paper/2005.00341/) (2020)
- [MusicLM: Generating Music From Text](https://mlsanity.com/paper/2301.11325/) (2023)
- [LTX-2: Efficient Joint Audio-Visual Foundation Model](https://mlsanity.com/paper/2601.03233/) (2026)
- [MuseCoco: Generating Symbolic Music from Text](https://mlsanity.com/paper/2306.00110/) (2023)
- [MMM : Exploring Conditional Multi-Track Music Generation with the Transformer](https://mlsanity.com/paper/2008.06048/) (2020)
- [MOVA: Towards Scalable and Synchronized Video-Audio Generation](https://mlsanity.com/paper/2602.08794/) (2026)
- [ComposerX: Multi-Agent Symbolic Music Composition with LLMs](https://mlsanity.com/paper/2404.18081/) (2024)
- [Foundation Models for Music: A Survey](https://mlsanity.com/paper/2408.14340/) (2024)
- [Anticipatory Music Transformer](https://mlsanity.com/paper/2306.08620/) (2023)
- [Frechet Music Distance: A Metric For Generative Symbolic Music Evaluation](https://mlsanity.com/paper/2412.07948/) (2024)
- [Feedforward 3D Editing Learns from Semantic-Part Transformation](https://mlsanity.com/paper/2605.27351/) (2026)
- [From Frames to Sequences: Temporally Consistent Human-Centric Dense Prediction](https://mlsanity.com/paper/2602.01661/) (2026)
- [A Comprehensive Survey for Evaluation Methodologies of AI-Generated Music](https://mlsanity.com/paper/2308.13736/) (2023)
- [Soulstyler: Using Large Language Model to Guide Image Style Transfer for Target Object](https://mlsanity.com/paper/2311.13562/) (2023)
- [One Video, One World: Turning Monocular Video into Physical 4D Scenes](https://mlsanity.com/paper/2606.31388/) (2026)
- [Engine-Native Editable 3D World Reconstruction with Objects and Lighting](https://mlsanity.com/paper/2607.20889/) (2026)
- [EmoGen: Eliminating Subjective Bias in Emotional Music Generation](https://mlsanity.com/paper/2307.01229/) (2023)
- [Emotion-driven Piano Music Generation via Two-stage Disentanglement and Functional Representation](https://mlsanity.com/paper/2407.20955/) (2024)
- [Qwen-Music Technical Report](https://mlsanity.com/paper/2607.11699/) (2026)
- [PairCoder++: Pair Programming as a Universal Paradigm for Verified Code-Driven Multimodal and Structured-Artifact Generation](https://mlsanity.com/paper/2607.01883/) (2026)
- [Pushing the Frontier of Full-Song Generation: Hierarchical Autoregressive Planning Meets Flow-Matching Rendering](https://mlsanity.com/paper/2607.20253/) (2026)
- [Exploring Tokenization Methods for Multitrack Sheet Music Generation](https://mlsanity.com/paper/2410.17584/) (2024)
- [SVGFusion: A VAE-Diffusion Transformer for Vector Graphic Generation](https://mlsanity.com/paper/2412.10437/) (2024)
- [BEAT: Tokenizing and Generating Symbolic Music by Uniform Temporal Steps](https://mlsanity.com/paper/2604.19532/) (2026)
- [Segment-Factorized Full-Song Generation on Symbolic Piano Music](https://mlsanity.com/paper/2510.05881/) (2025)
- [Text2Score: Generating Sheet Music From Textual Prompts](https://mlsanity.com/paper/2605.13431/) (2026)
- [ARIMA: Reconstruction-Grounded Predictive Representation Learning for Symbolic Music](https://mlsanity.com/paper/2607.10003/) (2026)
- [Efficient Long-Sequence Diffusion Modeling for Symbolic Music Generation](https://mlsanity.com/paper/2603.00576/) (2026)

## Abstract (arXiv)

Text-to-music language models begin with a choice usually made by default: how to tokenize music. Normally entangled with backbone, data, and recipe, its effect has never been measured in isolation. We fix pretrained Qwen3.5 (0.8B-27B), data, budget, and decoding, and swap only the representation across seven tokenizations, anchoring texture metrics to each representation's model-free ceiling. The ordering is clean and surprising: representation, not model size, is the binding variable for distributional fidelity. Scaling the backbone 34x barely moves Frechet Music Distance (FMD), whereas switching representation halves it. PMT, a performance-resolution stream we release (10 ms timing, per-note velocity, multi-track texture; 609 symbols), reaches FMD 159 at 0.8B against 272-286 for beat grids (1.7-1.8x lower, up to 2.8x elsewhere; non-overlapping bootstrap CIs), so a 0.8B performance-resolution model beats a 27B beat grid. It reappears on a 26M from-scratch backbone and a second performance-resolution tokenizer: a property of the class, not one lucky vocabulary. Nor is it a finer-lattice artifact: snapping PMT's onsets to the beat grids' resolution still leaves it 67-129 FMD ahead of both (n=500). The effect is distributional; whether it is audible is a separate question, left open by our probe, with a human study pre-registered. Native caption adherence is weak but separable: a lightweight decode-time constraint doubles instrument-F1 (.28 to .60) and Correct-Key (.16 to .35) at no distributional cost. We release the harness, 25+ checkpoints, two corpora (86.6k aligned across caption/MIDI/ABC/audio; 6.25M captioned, the largest for music), and an imprinting diagnostic: published text-to-MIDI systems reproduce their training distribution near-invariant to the caption (72% vs. 71% chord-time on disjoint domains). The field's next representation claim can now be measured, not asserted.