Gray Swan
PulseAugur coverage of Gray Swan — every cluster mentioning Gray Swan across labs, papers, and developer communities, ranked by signal.
- 2026-05-28 funding AI security startup Gray Swan raised $40 million in Series A funding, reaching a $200 million valuation. source
3 day(s) with sentiment data
-
Google DeepMind unveils Gemini 4 Carbon for coding and Genie 4 for world simulation
Google DeepMind has announced two new frontier AI models: Gemini 4 Carbon and Genie 4. Gemini 4 Carbon is designed for complex, long-horizon tasks, excelling in multi-day coding, cybersecurity defense, and bug patching,…
-
Google releases Gemini 4 Carbon, topping AI benchmarks
Google has released Gemini 4 Carbon, the latest and most advanced model in its Gemini 4 family, positioning it as a leader in long-form engineering tasks and cybersecurity. The model achieved the highest score on the Ar…
-
Top LLMs for Browser Agents in 2026 Ranked by Score and Cost
A comparative analysis of LLMs for browser agents in October 2026 highlights Claude Opus 5.5 as the top performer for success rate, with Claude Sonnet 5.5 offering a similar score at half the price. For budget-conscious…
-
Gray Swan relaunches AI red-teaming platform for AI agents
AI security firm Gray Swan has relaunched its red-teaming platform, Arena, to address the evolving security challenges posed by AI agents. The updated platform, Arena 2.0, focuses on new attack vectors like Indirect Pro…
-
Google launches Gemini 4 Argon, prioritizing cybersecurity partners
Google has launched Gemini 4 Argon, its latest advanced AI model, focusing on cybersecurity and complex coding tasks. The model is initially available to select cybersecurity partners through Google's Fairwind Program, …
-
Anthropic's Opus 5 shows major gains in prompt injection resistance · 1 source tracked
Anthropic's Opus 5 model demonstrates significantly improved resistance to prompt injection attacks, achieving a near-zero success rate when combined with additional system-level defenses. While the model itself is more…
-
Prompt injection attacks succeed by leveraging AI model functions, not overriding them
A recent exploration into AI red-teaming arenas revealed that direct commands to ignore previous instructions are ineffective against hardened models. Instead, successful prompt injection attacks leverage the model's in…
-
Autonomous agents risk double-purchasing due to system design flaws · 2 sources tracked
A critical vulnerability has been identified in autonomous agent commerce systems, where agents can inadvertently make duplicate purchases. This issue arises from a combination of factors: agents' inherent retry logic f…
-
Gray Swan tackles AI security risks with advanced red teaming tools
Gray Swan, a company co-founded by Zico Kolter and Matt Fredrikson, is developing tools and methodologies for AI security, focusing on "gray swan" events—unlikely but foreseeable AI incidents. Their work addresses vulne…
-
AI security firm Gray Swan raises $40M at $200M valuation
AI startup Gray Swan has secured $40 million in Series A funding, valuing the company at $200 million. The startup employs approximately 15,000 security professionals who act as hackers to pressure-test AI models from m…
-
OpenAI adds AI safety expert Zico Kolter to its Board of Directors
OpenAI has appointed Zico Kolter, a professor and head of Carnegie Mellon University's Machine Learning Department, to its Board of Directors. Kolter's expertise in AI safety, alignment, and model robustness is expected…