CoIn: Counting the Invisible Reasoning Tokens in Commercial Opaque LLM APIs

2025-05-19

AI Summary

Commercial closed-source LLM APIs employ invisible inference tokens—unobservable during API invocation—whose counts and semantic content are neither verifiable nor transparent, leading to opaque or even malicious billing inflation. To address this, we propose the first dual-dimensional verifiable auditing framework for LLM services. First, we construct a Merkle hash tree grounded in token-embedding fingerprints to enable cryptographic verification of token counts. Second, we integrate cosine similarity with semantic clustering to detect semantic inconsistencies and verify the authenticity of hidden tokens. Deployed as a trusted third-party auditor, our framework achieves up to 94.7% accuracy in identifying token-count inflation across real-world API calls. This significantly enhances billing transparency and strengthens user trust in commercial LLM services.

Technology Category

• Machine Learning: Large Multimodal Models (LMMs)
• Natural Language Processing: Fact-Checking / Misinformation Detection (NLP Focus)
• Computer Vision: Language and Vision

Application Category

• Economics, Online Markets and Human Computation: Cost models of using LLMs in production systems
• Search and Retrieval-Augmented AI: Search Tool Learning with LLM: Teaching LLMs to invoke search and make use of retrieved information
• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation

Abstract

As post-training techniques evolve, large language models (LLMs) are increasingly augmented with structured multi-step reasoning abilities, often optimized through reinforcement learning. These reasoning-enhanced models outperform standard LLMs on complex tasks and now underpin many commercial LLM APIs. However, to protect proprietary behavior and reduce verbosity, providers typically conceal the reasoning traces while returning only the final answer. This opacity introduces a critical transparency gap: users are billed for invisible reasoning tokens, which often account for the majority of the cost, yet have no means to verify their authenticity. This opens the door to token count inflation, where providers may overreport token usage or inject synthetic, low-effort tokens to inflate charges. To address this issue, we propose CoIn, a verification framework that audits both the quantity and semantic validity of hidden tokens. CoIn constructs a verifiable hash tree from token embedding fingerprints to check token counts, and uses embedding-based relevance matching to detect fabricated reasoning content. Experiments demonstrate that CoIn, when deployed as a trusted third-party auditor, can effectively detect token count inflation with a success rate reaching up to 94.7%, showing the strong ability to restore billing transparency in opaque LLM services. The dataset and code are available at https://github.com/CASE-Lab-UMD/LLM-Auditing-CoIn.

Problem

• Detects hidden reasoning tokens in commercial LLM APIs
• Verifies authenticity of billed but invisible tokens
• Prevents token count inflation by providers

Innovation

• Verifiable hash tree for token count auditing
• Embedding-based relevance matching for content validation
• Third-party auditing framework for billing transparency