Skip to content
VECTOR WIREAI INTELLIGENCE
UTC

AI Research Radar

November 7–13: 17 frontier papers out of 2,237 read.

17 frontier281 cleared2,237 papers read

Frontier

Papers rated as moving the field

How we choose

Singularity-Bench

Opus 5.5 with Claude Code reaches a Singularity-Index of 1.25, matching or beating human baselines on all six tasks, while remaining agents average 0.58.

Can Agents Build Agents?

AgentsNVIDIA, Amazon, MetaBenchmarkarXivPDF

Infrathon

Claude Opus 5.5 raises the mean KV-cache compression ratio from 6.82% with the human-designed Quant-VideoGen codec to 19.86%.

AgentsAllenAIBenchmark

Lazare, veni foras

Editing seven neurons turns Nemotron-Terminal-8B into a sleeper agent that fires the payload on at least 98.8% of triggered test trajectories.

SecurityMicrosoft

Bait

Bait mitigates harmful agentic reinforcement learning by probing the reward function while still allowing learning from benign rewards.

SecurityGoogleCode

When Experience Becomes a Backdoor

BadEvolve demonstrates effective task-level backdoor attacks on self-evolving LLM agents across diverse frameworks, benchmarks, and backbone LLMs.

SecurityMicrosoftCode

ManiLadder

ManiLadder benchmark with 112 tasks across 4 difficulty levels calibrates manipulation task difficulty using a unified imitation learning policy.

RoboticsBenchmark

Frontier3DBench

GPT-6.1-sol achieves the highest overall score of 66.02/100 on Frontier3DBench, while all 11 configurations struggle with reference reconstruction.

AgentsBenchmark

Reins

Reins lets GPT-6.1 Sol complete 88% of real dual-arm manipulation episodes on five tasks with no task-specific code.

Robotics

Jais 3

Jais 3 matches or outperforms open-weight models of comparable size on Arabic and agentic benchmarks with fewer hallucinations.

Agents

Compass 1

Compass 1 achieves a 79.4% overall success rate in open-world embodied navigation, versus 17.8% for the strongest baseline.

Robotics

Shortcuts and Boundaries

A binary compliance penalty raises safe completion on 32 unseen changing-evidence tasks from 53.1% to 89.1% while both policies achieve full real work.

Safety

Meituan-Robotics-0

Meituan-Robotics-0 achieves 93.6% success on RoboTwin 2.0 and outperforms π0.5 on three real-robot tasks.

Robotics

The week by field

Nov 7–13: 281 cleared, 17 frontier

Also cleared

264 papers, by field

Security & privacy94
All 94 papers in Security & privacy
Safety & alignment34
All 34 papers in Safety & alignment
Agents30
All 30 papers in Agents
Inference & efficiency28
All 28 papers in Inference & efficiency
AI & society21
All 21 papers in AI & society
Robotics & embodied AI15
All 15 papers in Robotics & embodied AI
Language & speech13
All 13 papers in Language & speech
Vision & multimodal11
All 11 papers in Vision & multimodal
Science & health7
All 7 papers in Science & health
Reasoning & math4
Machine learning4
Training & data3