Metraipolis › archive › dc_0f53dbd6eabb
The Reasoning Tax: Token Economics of LLM Reasoning Across Task Types and Deployment Contexts — arXiv:2608.26235
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2608.26235
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-17T14:22:27.789Z
- SHA-256 of the archived text
48fd51b238102135d92e8c4676283d08f2890a70917726cc363034db74a433fc- Stable URL of this page
https://metraipolis.com/documento/dc_0f53dbd6eabb
How to cite
The Reasoning Tax: Token Economics of LLM Reasoning Across Task Types and Deployment Contexts — arXiv:2608.26235 — arXiv (en), https://arxiv.org/abs/2608.26235 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_0f53dbd6eabb (Metraipolis archive dc_0f53dbd6eabb, sha256 48fd51b238102135d92e8c4676283d08f2890a70917726cc363034db74a433fc)
BibTeX
@misc{dc0f53dbd6eabb,
title = {The Reasoning Tax: Token Economics of LLM Reasoning Across Task Types and Deployment Contexts — arXiv:2608.26235},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2608.26235},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_0f53dbd6eabb; sha256 48fd51b238102135d92e8c4676283d08f2890a70917726cc363034db74a433fc},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2608.26235. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Accuracy-only benchmarking of reasoning-capable large language models misses a central deployment question: when do extended thinking tokens earn their cost? We introduce the Token Economy Score (TES), a marginal benchmarking metric that measures the accuracy gain of a reasoning model over a non-reasoning baseline, normalized by the generated-token multiplier. We define paired and approximated TES variants for model families with reasoning toggles and frontier models without direct non-reasoning counterparts. We then conduct an empirical benchmarking analysis across 151 model-benchmark evaluation runs on seven benchmarks spanning mathematics, code generation, science reasoning, instruction following, expert knowledge, knowledge recall, and research-level physics. The analysis examines three deployment-facing dimensions: which task structures yield positive marginal reasoning efficiency, how increasing reasoning effort changes TES within model families, and how deployment context changes economic viability. Results show that task structure predicts reasoning efficiency better than nominal difficulty: sequential inferencechain tasks such as AIME 2025 and LiveCodeBench show high TES, while knowledge-recall tasks such as MMLU-Pro show low TES despite their difficulty. We also find systematic diminishing returns at higher reasoning effort levels, including cases where additional thinking reduces accuracy. Finally, Reasoning Cost Share (RCS) shows that inference spend is often dominated by internal thinking, while Deployment Cost Multiplier (DCM) shows how on-premises deployment can change the economics of otherwise costly reasoning workloads. These findings support a benchmarking-driven model-selection rule: enable reasoning selectively by task type, effort level, and deployment context rather than treating it as a universally beneficial mode. Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-08-28 in cs.AI). Categorie arXiv: Artificial Intelligence (cs.AI) ; Performance (cs.PF). Pagina del lavoro: https://arxiv.org/abs/2608.26235 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito). Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina. Letto il 2026-10-17T13:16:53.476Z dalla pagina https://arxiv.org/catchup/cs.AI/2026-08-28?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).