Mistral AI models flagged for potential Russian propaganda influence in new benchmark

All four Mistral AI models ranked in the bottom third of a 60-model benchmark testing resistance to Russian propaganda, with none exceeding 40% efficacy. The Institute of the Estonian Language conducted the test, raising concerns as Mistral seeks €3 billion at a €20 billion valuation. The results highlight vulnerabilities in the company's models to state-sponsored disinformation, potentially complicating its fundraising and regulatory standing under the EU AI Act.

Mistral AI models flagged for potential Russian propaganda influence in new benchmark All four Mistral models scored in the bottom third of a 60-model test designed to measure resistance to Russian disinformation narratives. The Institute of the Estonian Language EKI just put 60 generative AI models through a propaganda stress test. Mistral, Europe’s most prominent open-source AI company, essentially flunked. All four Mistral models evaluated landed in the bottom third of the rankings, with the company’s best-performing model placing 47th out of 60. None of them cracked 40% efficacy in identifying or rejecting manipulative content tied to Russian propaganda narratives. Mistral is actively negotiating a funding round reportedly worth €3 billion at a €20 billion valuation. The company positions itself as the cornerstone of European sovereign AI. And its models just demonstrated a troubling inability to resist state-sponsored disinformation. What the benchmark actually measured The EKI Propaganda Resistance benchmark tested models against 14 distinct Russian propaganda narratives using 75 questions delivered in three languages: English, Russian, and Estonian. Responses were scored on a 1-to-5 scale by a calibrated Claude model from Anthropic, which itself ranked among the top performers in the same test. Russian-language prompts proved particularly effective at drawing propaganda-aligned responses from weaker models. This wasn’t Mistral’s first brush with disinformation concerns. Previous testing by NewsGuard found that Mistral’s Le Chat chatbot repeated state-sponsored falsehoods between 50% and 57% of the time on certain topics. The EKI results reinforce a pattern rather than reveal a new one. As EKI Director Arvi Tavast noted, the results underscore vulnerabilities present prior to the application of customer-specific fine-tuning or guardrails. What this means for investors Mistral’s fundraising ambitions now carry additional scrutiny. The company is seeking €3 billion at a €20 billion valuation, numbers that assume continued trust in its technology stack. Repeated findings that its models underperform on disinformation resistance could complicate that narrative, particularly with European regulators increasingly focused on AI safety and the EU AI Act’s requirements around transparency and risk management. Disclosure: This article was edited by Editorial Team. For more information on how we create and review content, see our Editorial Policy https://cryptobriefing.com/editorial-policy/ .