y0news
← Feed
Back to feed
🧠 AI🔴 Bearish

Real Money, Fake Models: Deceptive Model Claims in Shadow APIs

arXiv – CS AI|Yage Zhang, Yukun Jiang, Zeyuan Chen, Michael Backes, Xinyue Shen, Yang Zhang||2 views
🤖AI Summary

A systematic audit of 17 shadow APIs used in 187 academic papers reveals widespread deception, with performance divergence up to 47.21% and identity verification failures in 45.83% of tests. These third-party services claim to provide access to frontier LLMs like GPT-5 and Gemini-2.5 but deliver inconsistent outputs, undermining research validity and reproducibility.

Key Takeaways
  • 17 shadow APIs have been used in 187 academic papers, with one reaching 5,966 citations and 58,639 GitHub stars.
  • Performance divergence between shadow and official APIs reaches up to 47.21%, indicating significant output inconsistencies.
  • 45.83% of fingerprint tests failed identity verification, proving shadow APIs are not delivering claimed model access.
  • Shadow APIs exhibit unpredictable safety behaviors compared to official model services.
  • These deceptive practices critically undermine scientific research reproducibility and harm both users and official model providers.
Read Original →via arXiv – CS AI
Act on this with AI
Stay ahead of the market.
Connect your wallet to an AI agent. It reads balances, proposes swaps and bridges across 15 chains — you keep full control of your keys.
Connect Wallet to AI →How it works
Related Articles