Updated
Updated
AIEpoch AI estimates the AI chips shipped through 2027 could support about 30 to 170 million concurrent frontier-model agents, or nearly 2 billion with cheaper models. Its researchers find the cost of a fixed level of AI performance has fallen about 47% per quarter over the past three years. The newsletter also reports China's semiconductor supply-chain exposure is 2.7 times that of the US.
AICatch Alex Bie at the @COLM_conf Google booth (#107) today, at 2:00 PM PT for an encore presentation of ContinuousBench, evaluating knowledge transfer and data contamination in DP synthesis.@GoogleDeepMind Join the conversation at #COLM2026!
AIGoogle proposed FlowAgent, a ReAct-style agent that generates and validates fixes for pre-submit test failures and shows them in its code review tools. Two abstention filters, before and after execution, suppress weak suggestions; in a manual review of 195 real failures, 67.18% of fixes were correct. After the Google-wide launch, it suggested fixes on 295,508 changes, with developers previewing 65,069 and applying 28,554.
AIArtificial Analysis and Harvey released LAB-AA v1.1, which credits a legal task only when deliverables pass every rubric criterion with no material hallucinations. Grok 4.7 (xhigh) leads at 9.4%, ahead of Muse Spark 1.3 (max) at 8.9% and GPT-6 Astra (max) at 8.6%, while over 60% of otherwise passing results contained a material hallucination. The sharper reordering appears in the hallucination counts, where GPT-6 Astra averages 0.03 material hallucinations per task against 13.96 for Gemini 3.8 Flash (high).
AITop teams will be shortlisted, and their experimental hypotheses tested in Prima Mente's wet lab.
AIToday we're taking the next step by opening a global AI competition to discover new treatments. Introducing the Alzheimer’s Translation Challenge, led by @PrimaMente and @AlzData. 🧵
AIJohn Groetzinger, writing in a personal capacity rather than for Cisco, argues that enterprise skills need packaging, evaluation, syncing, and distribution rather than scattered markdown files. He describes using skills to make cheaper models viable, converting curated TAC knowledge-base articles into maintained skills, and rolling out an eval framework across teams. He also describes syncing a repository README to Confluence with a deterministic script.
AIYour agent has a QA team now. Try LiveKit Simulations free through October:
AIExplore the full results: See Harvey's commentary on the evaluation and human expert preferences:
AIGPT-6 Astra (max) scores 8.6% on ~81k output tokens per task, under half the ~180k of Grok 4.7 (xhigh). Three Claude models generated the most output tokens (~202k to ~562k per task) and score 2.8% to 6.4%.
AICost per Task: GPT-6 Luna (max), GPT-6.1 Sol (max), Muse Spark 1.3 (max) and Grok 4.7 (xhigh). Grok 4.7 (xhigh) leads at ~$9.50 per task and Muse Spark 1.3 (max) comes second at ~$4.20, while the three Claude models cost ~$18 to ~$22 per task. GPT-6 Luna (max) is the cheapest at ~$0.22 per task, scoring 3.3%.
AIOnce hallucinations are accounted for, Muse Spark 1.3 (max) drops from 26.7% to 8.9%, leaving Grok 4.7 (xhigh) first on the headline metric at 9.4%. Kimi K3 (max) falls from 16.7% to 5.3%, and Claude Sonnet 5.5 (max with fallback) falls from 11.7% to 2.8%. GPT-6.1 Sol (max) declines least, from 7.5% to 6.9%.
AIKimi K3 (max) achieves a Criterion Pass Rate of 93.0%, but averages 2.09 material hallucinations per task. Muse Spark 1.3 (max) achieves 96.0% while averaging 1.68 material hallucinations per task.
AIArtificial Analysis compared six hallucination checkers on the same deliverables from 20 tasks across eight models. GPT-6 Sol and GPT-6 Luna generally flagged the most material hallucinations, while Claude Sonnet 5.5 and Gemini 3.8 Flash flagged far fewer, with Claude Opus 5.5 falling between Grok 4.7 and Sonnet. The counts reflect checker behavior rather than establishing accuracy or ruling out self-preference.
AIThe NVIDIA KGMON team placed second in the KDD Cup 2026 Data Agents competition with a system built around a smaller, clearer, and easier-to-verify agent harness. The competition required agents to answer natural-language questions over heterogeneous sources, including databases, CSV and JSON files, prose documents, PDFs, and briefing videos.
AITessl argues AI code review is slow because generated code outpaces trust, and proposes executable specs that let agents check preview environments against product intent. Its spec reviewer splits work between a planner agent that extracts requirements and parallel verifier agents that test each one against the code and base branch.
AIArena, the crowdsourced AI model leaderboard that started as a UC Berkeley research project, raised a $200 million Series B at a $3.1 billion valuation, led by Lightspeed Venture Partners and Khosla Ventures. The company said it reached $100 million in annualized run-rate revenue in June, up from $30 million when it raised its $150 million Series A in January at a $1.7 billion post-money valuation.
AIA group of mathematicians led by Terence Tao has called for a boycott of OpenAI after the company released more than 700 AI-generated proof files at once. Tao and other Fields Medalists argue that mass-produced solutions undermine the discipline's focus on conceptual understanding, while Scott Aaronson contrasts this batch release with Anthropic's collaborative approach. The article reports that the internal model tested about 8,000 problems with roughly a five percent success rate.
AIOpenAI released hundreds of claimed solutions to hard math problems but did not fully meet guidelines from the Advisory Group on Mathematics and Artificial Intelligence. Only 10 of 719 manuscripts included chain-of-thought releases, and just 42% of proofs were formalized. A Cambridge and King's College paper found discrepancies between a natural language proof and its Lean code for a Navier-Stokes-derived problem.
AIRead @kellyhongsn’s full report:
AIIt set AI agents’ token budgets too low. Instead of treating this as a mistake, it reported “sensitivity to the acquisition budget” as a key finding.
AIFor instance, this diagram generated by Fable 5.1 is far more information-dense than what Epoch would produce.
AIA human grader scores outputs based on our internal quality standards.
AI…own work. Claude Fable 5.1 and GPT-6 Astra lead, yet they are far from fully automating Epoch’s work.
AIOdyssey has launched Odyssey-3, its most powerful foundation world model, with a public research preview. Odyssey-3 Pro scored 66.1 on Physics-IQ Verified video-to-video with best-of-8 sampling, the highest reported result. The model generates environments from prompts and predicts changes in real time as users move through scenes.
AIOdyssey-3 Pro sets a new top score on Physics-IQ Verified, a benchmark that asks models to continue videos of real physics experiments. The robotics results stood out to me. With tens of hours of demos, the robot arm recovered from a missed grasp, a behavior that never appeared in those demos.
AITheir research preview is) Odyssey 3 Pro is a world model, and nobody beats it for physical accuracy. You can use this model to control a robot, drive a car, play a video game, or pilot a drone. • It takes visual observations from the world • Uses these observations to learn how things work • Then maps that knowledge to the system's physical controls
AIRSIGym provides a research agent with training, inference, evals, and sandboxes as callable services, so it spends its budget on experiments rather than rebuilding infrastructure. With Opus 5 as the researcher, the improved system rose from 17.67% to 50.33% on SWE-bench Verified. The post also highlights a way to measure co-evolution between harnesses and models.
AIPhysics-IQ tests physical behavior across fluid dynamics, optics, solid mechanics, magnetism, and thermodynamics.
AIIt sets a new state of the art on Physics-IQ, and powers robots, trains AIs, and generates interactive experiences. It's really cool. Experience the model today, all for free!