# The Price Reversal Phenomenon: When Cheaper Reasoning Models Cost More

arXiv: https://arxiv.org/abs/2603.23971 · HTML: https://mlsanity.com/paper/2603.23971/

## Key facts

- Cited by 4 papers in the map (data through 14 Oct 2026).
- Placed in Test-time compute and efficient thinking, one of 17 directions in LLM reasoning.
- Builds on 5 of its 56 references that are in the map.
- First posted on arXiv 25 Mar 2026.

## Placed in

- LLM reasoning / [Test-time compute and efficient thinking](https://mlsanity.com/lr/test-time-compute-and-efficient-thinking/)

## Builds on (most cited first, 5 of 5)

- [AutoGen: Enabling Next-Gen LLM Applications via Multi-Agent Conversation](https://mlsanity.com/paper/2308.08155/) (2023)
- [Humanity's Last Exam](https://mlsanity.com/paper/2501.14249/) (2025)
- [Towards Optimizing the Costs of LLM Usage](https://mlsanity.com/paper/2402.01742/) (2024)
- [Optimising Calls to Large Language Models with Uncertainty-Based Two-Tier Selection](https://mlsanity.com/paper/2405.02134/) (2024)
- [Boulder2Vec: Modeling Climber Performances in Professional Bouldering Competitions](https://mlsanity.com/paper/2411.02343/) (2024)

## Built on by (most cited first, 4 of 4)

- [How Do AI Agents Spend Your Money? Analyzing and Predicting Token Consumption in Agentic Coding Tasks](https://mlsanity.com/paper/2604.22750/) (2026)
- [Statistically-Lossless Quantization of Large Language Models](https://mlsanity.com/paper/2605.02404/) (2026)
- [Can a Lightweight Multimodal Model Estimate LLM Reasoning Performance? A Study for Compute-Optimal Document Inference](https://mlsanity.com/paper/2608.18591/) (2026)
- [Learning the Cost of Reliable Inference](https://mlsanity.com/paper/2609.28322/) (2026)

## Abstract (arXiv)

Developers and consumers increasingly choose reasoning models (RMs) based on their listed API prices. However, how accurately do these prices reflect actual inference costs? We conduct the first systematic study of this question, evaluating 8 frontier RMs across 12 diverse tasks covering competition math, science QA, code generation, and multi-domain agents. We uncover the pricing reversal phenomenon: in 32% of model-pair comparisons, the model with a lower listed price actually incurs a higher total cost, with reversal magnitude reaching up to 28x. For example, Gemini 3 Flash's listed price is 80% cheaper than GPT-5.4's, yet its actual cost across all tasks is 38% higher. We build a formal cost attribution framework based on Shapley value, and leverage it to trace the dominating contributors to vast heterogeneity in thinking token consumption and number of interaction turns: on the same query, one model may use 900% more thinking tokens than another, or 10x more turns of environment interactions. We further show that per-query cost prediction is fundamentally difficult: repeated runs of the same query yield thinking token variation up to 9.7x, establishing an irreducible noise floor for any predictor. Thus, we propose cost distribution prediction as an open challenge. Our findings demonstrate that listed API pricing is an unreliable proxy for actual cost, calling for cost-aware model selection and transparent per-request cost monitoring.