🤖 AI Agents Weekly: Gemini 4 Carbon, GPT-6.1 Terra, OpenAI's Three Discoveries, Mistral Large 4, Axiom CFSG Proof, Qwen Investigation, Genie 4, and More

Google DeepMind previewed Gemini 4 Carbon, a frontier model that ranks first on the Artificial Analysis Intelligence Index v5 at 59, ahead of Claude Opus 5.5 at 58 and GPT-6 Astra at 57, with Google reporting 65.2% on Terminal-Bench 4.0 and 77.4% on Agents' Last Exam v1.1. Google said internal teams have run Gemini 4 Carbon for days on multi-step engineering work such as root-causing and patching bugs in large codebases, and reported 0.0% successful cheating on Terminal-Bench 4.0, almost 80% fewer high-severity misalignment alerts in internal live coding environments, and a 0.0% attack success rate on Gray Swan prompt injection tests. Access begins with a subset of cyber defenders through the Fairwind Program at introductory pricing of $2 per million input tokens and $10 per million output tokens, rising to $4 and $20, with a 1M-token context window.

In today’s issue: - Google previews Gemini 4 Carbon - OpenAI ships GPT-6.1 Terra - OpenAI models refute Exponential Time Hypothesis - Mistral opens Large 4 weights - Agent swarm formalizes 15,000-page proof - Cohere flags Qwen upload attempt - Anthropic measures agent influence on experts - Cognition launches FrontierCode 2.0 - DeepMind unveils Genie 4 - Pretraining poison backdoors CLI agents And all the top AI dev news, papers, and tools. Top Stories Gemini 4 Carbon Google DeepMind introduced Gemini 4 Carbon, its next frontier model, with state-of-the-art cybersecurity and long-horizon capabilities. Google says internal teams have seen it run for days on multi-step engineering work such as root-causing and patching bugs in large codebases. - Benchmarks: Carbon ranks first on the Artificial Analysis Intelligence Index v5 at 59, ahead of Claude Opus 5.5 58 and GPT-6 Astra 57 , and Google’s comparison table shows 65.2% on Terminal-Bench 4.0 and 77.4% on Agents’ Last Exam v1.1, ahead of GPT-6 Astra and Claude Opus 5.5. Cognition’s FrontierCode 2.0 puts it at 40%, behind Claude Opus 5.5, GPT-6 Astra, and Claude Sonnet 5.5. - Cybersecurity: Evaluated without guardrails, it reaches 88.2% on SEC-bench Pro for finding vulnerabilities, 54.8% on ExploitGym for exploitation, and 71.0% on CWE-bench for patching. - Alignment: Google reports successful cheating on Terminal-Bench 4.0 at 0.0%, almost 80% fewer high-severity misalignment alerts in internal live coding environments, and a 0.0% attack success rate on Gray Swan prompt injection tests. - Cost: Artificial Analysis measures about $7.76 per Index task, roughly 23% less than Claude Opus 5.5 at max effort, with a 1M-token context window. - Availability: Access starts with a subset of cyber defenders through the Fairwind Program, and Google says developers, businesses, and consumers will follow. Pricing is $2 input and $10 output per million tokens during the introductory period, then $4 and $20.