Stealing Gemini 2.5's Hidden Architecture via Streaming API Side-Channel
alexisjross · x · 2026-08-27
Researchers found a way to steal hidden model architecture and undocumented inference optimizations from production LLMs through ordinary streaming APIs. Tests on Gemini Flash 2.5 revealed:
- Latency Spike: A 3.2× latency jump at 130k tokens exposed the use of speculative decoding.
- Context Window: A hidden 128K draft-model context window was discovered.
This method leverages timing characteristics during inference to reveal internal implementation details not disclosed by the vendor.
Related event: Researchers steal hidden LLM architecture via streaming API timing(3 posts)→
More from Safety
- Cisco firewall console bug CVE-2026-20079 grants unauthenticated root access — victorialslocum · 2026-10-15
- ChatGPT ads can target you by phone number, email, and your prompts — 404 Media · 2026-10-15
- Anthropic's Mythos 5.1 pledges to treat Claude's expressed disagreement as decision input — Hesamation · 2026-10-15
- Stolen Thoughts: stealing hidden reasoning traces from LLM APIs in two calls — maksym_andr · 2026-10-15
- Report: US Treasury Pressuring Countries to Adopt 'Super Intelligence' Over AI — Polymarket · 2026-10-15
- ServiceNow launches AI Passport to give AI agents verified identities, with NVIDIA as first partner — azavery · 2026-10-15