OpenAI / ChatGPT

Meta has released Muse Spark 1.3, their fourth Muse Spark model release in five months. Muse Spark 1.3 (max), which is in limited preview for Meta’s partners, scores 62 on the Artificial Analysis Intelligence Index, behind only Claude Fable 5.1 and Claude Opus 5. The variant available now, Muse Spark 1.3 (xhigh), scores 61 and ties with GPT-5.6 Sol (max) and Grok 4.6 (high). Both variants’ gains come primarily from improvements in agentic work and scientific capabilities Muse Spark 1.3 (xhigh) enters the Artificial Analysis Intelligence Index at 61, up 4 points from Muse Spark 1.2 (57, August) and 8 points from Muse Spark 1.1 (53, July). It enters tied with GPT-5.6 Sol (max), Grok 4.6 (high), and Claude Opus 5 (high), and behind Claude Fable 5.1 (max, 66), Claude Opus 5 (max, 63), and Claude Fable 5 (max, 62) Muse Spark 1.3 (max), which is in a limited preview stage, lands at 62. This higher index score is enabled by gains vs. Muse Spark 1.3 (xhigh) in Tau3-Bench Banking (52% vs. 47%) and GDPval-AA v2 (1,754 Elo vs. 1,709). Muse Spark 1.3 (max) is second only to Claude’s Fable and Opus variants in total score Congratulations to @AIatMeta, @finkd, and @alexandr_wang on the release! Key Takeaways: ➤ Continued improvement on agentic knowledge work tasks. At the launch of Muse Spark 1.2, we noted its significant gains in agentic knowledge work performance vs. Muse Spark 1.1. The latest iteration continues this trend, with Muse Spark 1.3 (xhigh) demonstrating a notable 12-point gain vs. Muse Spark 1.2 in Tau3-Bench Banking (35% to 47%), a 5-point gain in Terminal-Bench 2.1 (80% to 85%), and a new GDPval-AA v2 Elo of 1709 against its predecessor’s 1615. Muse Spark 1.3 (max) improves further on Tau3-Bench Banking (52%) and GDPval-AA v2 (1,754 Elo). This Tau3-Bench Banking score is #1 among all models. Muse Spark 1.3 (max) achieves these higher agentic work scores by using more turns and total reasoning tokens, reasoning 62% more on GDPval-AA v2 and 28% more on Tau3-Bench Banking compared to Muse Spark 1.3 (xhigh) ➤ The lowest cost per task for any model at 59+ on the Artificial Analysis Intelligence Index. Muse Spark 1.3 (xhigh) costs $0.55 per Intelligence Index task at Meta's unchanged $1.25/$4.25 per 1M token pricing ($0.15 for cached input), with its peers GPT-5.6 Sol (max) and Grok 4.6 (high) costing $0.95 and $0.94 respectively, a 70%+ premium. This places Muse Spark 1.3 (xhigh) on the Pareto frontier for Intelligence vs. Cost per Task. Its cost per task is higher than Muse Spark 1.2 ($0.40 per task), driven by ~57% more input tokens per task on agentic evaluations, with output tokens up only ~8%. Pricing for Muse Spark 1.3 (max) is not yet publicly available ➤ Scientific Reasoning results rose across the board, led by CritPt. CritPt was the standout non-agentic score gain vs. Muse Spark 1.2, with a material +8 points for the xhigh variant (18% to 26%), and GPQA Diamond achieved +4 points (90% to 94%), while Humanity’s Last Exam and SciCode each gained a more modest 2-3 points (45% to 47% and 56% to 59%, respectively). Muse Spark 1.3 (max) achieved roughly similar scores to the xhigh variant, gaining 2 points in Humanity’s Last Exam, tying on GPQA Diamond, and losing a point on CritPt vs. Muse Spark 1.3 (xhigh) ➤ Minor regressions in only two evaluations. Both Muse Spark 1.3 (xhigh) and Muse Spark 1.3 (max) dropped 4 points in AA-LCR (83% to 79%) when compared to Muse Spark 1.2, and AA-Omniscience (Accuracy) fell 3 points for xhigh and 1 point for max. The drops in AA-Omniscience (Accuracy) are due to a higher abstention rate (not answering questions when unsure), which also lowered the hallucination rate for Muse Spark 1.3 (xhigh) Other model details (xhigh variant): ➤ Context window: 1M tokens, unchanged from Muse Spark 1.2 ➤ Pricing: unchanged from Muse Spark 1.2: $1.25/$4.25 per 1M input/output tokens, with cache hits discounted to $0.15 per 1M ➤ Input modalities: text, image, video ➤ Availability: Meta's first-party API and Muse Code

The Trump administration has now formally put the U.S. government behind OpenAI’s core fair-use argument in its copyright fight with The New York Times. some of the most conclusive statements they said in their filed document. - “the ‘training of AI models on copyrighted material,’ in and of itself, ‘does not violate copyright laws.’” - “For all these reasons, the United States has a strong interest in this Court rejecting any argument that training LLMs on copyrighted texts violates copyright law.” - “The fourth fair use factor … supports the conclusion that OpenAI’s model training using New York Times articles is fair use.” - “The copying of protected text articles as part of training an LLM is a use of a different kind or character that is ‘transformative—spectacularly so.’” - “In sum, the use of copies to train LLMs is extraordinarily transformative.” “Rules of law that make it significantly more difficult to develop a robust AI industry in the United States therefore threaten national security and give a competitive advantage to foreign adversaries who are not so encumbered.” - “In this litigation, the New York Times seeks to narrow fair-use doctrine to exclude the training of OpenAI’s large language models (LLMs). That result would be inconsistent with basic copyright law principles and severely hamper ‘the Progress of Science and useful Arts.’” “But it would be problematic—and legally incorrect—to impose broad copyright liability that would generally render training of AI models impermissible without licensing.”

A massive win for OpenAI (and for AI training in general) for its legal case against New York Times. The U.S. govt just formally backed OpenAI’s claim that copyrighted-text training is fair use, partly on national-security grounds. This is Washington’s first formal intervention in the wider wave of copyright lawsuits over AI training, though the filing is advisory rather than binding on the court. The U.S. Dept of Justice filed a statement of interest of the US formally arguing in the OpenAI copyright litigation that training LLMs on copyrighted texts should generally qualify as fair use; That distinction still leaves separate copyright questions around how training data was acquired and whether particular outputs reproduce protected passages. The Justice dept separated acquiring material, training on it, and generating outputs, then focuses its argument specifically on copying at the training stage. It argues that training serves a different purpose from publishing an article because an LLM uses text to learn statistical relationships and generate new responses. For market harm, DOJ says training itself does not substitute for the original work, so later AI-generated competition should not automatically make the earlier training unlawful. The administration also warns that blanket licensing requirements could raise barriers for smaller AI companies and put U.S. developers at a disadvantage against foreign competitors. The court must still decide fair use case by case, but adopting DOJ’s framework would shift much of the legal pressure from model training toward data acquisition and specific outputs.

OpenAI says Astra is its first model to reach the Critical cybersecurity capability threshold. Under its Preparedness Framework, that means Astra can, with the right tools and access, find unknown flaws and develop exploits across hardened systems without step-by-step human guidance. Hence, OpenAI now says Astra will launch with additional chain-of-thought monitoring, while classifiers can automatically stop potentially unauthorized actions. Astra reached roughly 39% exploit success at ~75K output tokens, while GPT-5.6 Sol is only around 1% there and needs nearly 140K tokens to reach ~12% i.e. Astra is dramatically more capable and token-efficient at exploit development on this internal benchmark. Expert assessments went further: Astra escaped a browser sandbox, executed host commands, and escalated an unprivileged operating-system user to root.

As we prepare to release Astra, we’re focused on making increasingly capable AI safe and broadly accessible. Astra represents a significant advance in cybersecurity capability, reaching the Critical threshold under our Preparedness Framework. We're previewing how we evaluated the model, how its safeguards have advanced alongside its capabilities, and what we'll continue to learn and improve. https://t.co/OrrTgdU90K