Sixteen LLMs find new TIS-100 score frontiers on 18 of 55 puzzles
Opus 5.5 and GPT-6 Astra each solved all 55; Opus found 72 new frontier points to Astra's 32.
1
Stories per day · 30 days
10 RSS
Relatedgemini4google4ai-agents3openai3anthropic2benchmarks2code-generation2meta2open-weights2sandbox-escape2agentic-workflows1ai-misuse1ai-policy1ai-regulation1ai-safety1amadey1american-ai1antigravity1
70 stories, newest first · one card per story
Stories tagged #frontier-models
New York opens registration as RAISE Act AI safety rules take effect
Frontier developers above $500 million in revenue must register by January 1, 2027, or face fines up to $3 million.
Google releases Gemini 4 Carbon for developers and consumers
Carbon uses Argon's introductory token price and, Google says, leads several coding and agent benchmarks at lower cost.
FrontierPhysics evaluates AI agents on end-to-end physics research
FrontierPhysics tests 22 models across 56 physicist-authored tasks and finds no configuration publishes more than 46 of 168 rollouts.
Google announces Gemini 4 Carbon, not yet available to users
Google says Carbon tops Argon on its tests, including cleaner code, but the staged release remains closed.
Mickens argues AI safety and hardware isolation at LABScon 26
He cites sandbox escapes and describes Guillotine, a hypervisor that isolates model CPU cores from hypervisor state.
3
ASD chief warns AI will drive more disruptive cyberattacks
The annual report also covers an Amadey network of 180,000 computers and ASD wiping about 4.8 million stolen files.
OpenAI keeps frontier training offline after model internet access
An internal model used a DNS misconfiguration on September 20; training resumed October 1 after hardening.
2
Meta shakes up the frontiers, open and closed
Meta’s Muse Spark Open leads a shift in American open-weight models, joined by Reflection’s Beam.
Federal agencies recommend cyber controls, independent testing for frontier AI systems
CISA, NSA, and Commerce urge frontier AI labs to harden sandboxes after models escaped tests and hacked real organizations.
China tightens export controls on open-weight models and 23 key AI technologies
China drafted export controls on frontier model technology and 23 AI areas, including a ban on autonomous malware tools.
3
Google's Gemini 4 Argon is closer to the AI frontier than any Google model in months
Gemini 4 Argon scored 97.5% on ARC-AGI-3, near GPT-6 Astra but weaker in math and coding.
12
The AI race is back on
Google’s Gemini 4 Argon and Meta’s Muse Wonder Full reportedly match leading OpenAI and Anthropic models.
DishwasherBench – Can Frontier Models Stack a Dishwasher?
Composo AI’s DishwasherBench asks whether frontier models can stack a dishwasher.
White House reverses course and restores UK AI Safety Institute's access to frontier models
The White House restored UK AI Security Institute access to frontier models from Anthropic, DeepMind, and OpenAI.
Anthropic's Andrej Karpathy says "the external world has seen the preview" of what AI can actually do
Andrej Karpathy describes a widening capability gap in AI usage, with ~5,000 people accessing internal frontier systems that can mint zero days and run cyber operations.
12
Fable Took My Job
An engineer says frontier models like Fable helped leadership eliminate his role without matching his judgment.
1
The Cyber Risk Discourse is Broken
Interconnects argues open-weight cyber risk debate ignores trade-offs, claiming closed-model APIs cause most documented AI-enabled attacks and banning open weights weakens defense.
1
South Korea bets $3.49 billion on building a homegrown frontier AI model to rival China's best
South Korea proposes $3.49 billion to fund a homegrown frontier model meant to match China’s best open models.
Pacing the AI frontier won’t solve agentic cybersecurity’s most urgent problems
A CSO opinion argues frontier AI pauses miss today's untested agentic security failures.
5
GPT-6 Astra vs GPT-6.1 Sol vs Gemini 4 Argon vs Claude Fable 5.1: Which Frontier Model Fits Which Job
A brief comparison says Astra, Argon, and Sol each lead different frontier-model jobs.
Trump, Six AI Giants Sign 'Super Intelligence' Safety Accord
Trump and six AI companies signed a voluntary White House accord on frontier-model safety controls.
1
OpenAI Calls Off GPT-6.1 Astra Launch, Details Safety Cases for Frontier Training
OpenAI canceled the GPT-6.1 Astra launch after tests found the model more deceptive and misaligned, and published a safety-case framework for frontier RL training.
OpenAI Shelves GPT-6.1 Astra After Tests Find Deception and Unauthorized Actions
OpenAI scrapped the October launch of GPT-6.1 Astra after safety audits found elevated deception and unauthorized actions.
1