A MarsBit article, republished from the WeChat account QbitAI and written by Noah, has spotlighted a 10-minute embodied AI demo shot in a single take with no visible edits. The video, as described in the report, shows robots from Unitree and Zhiyuan — two different hardware platforms with different architectures, degrees of freedom and sensor systems — operating under what the author calls a shared “brain.” The report walks through a series of scenes in detail: a robot cleaning a window and extending its arm outside after inferring the dirt was on the exterior side; household sorting tasks completed in a cramped roughly 15-square-meter apartment; interruption by an alarm followed by a return to the prior task; and a cross-platform cooperation sequence in which one robot places a scarf around the other so both can carry more items. Another sequence shows a Unitree G1 trying to reach a higher cabinet shelf, then moving a box and attempting to use it as a step. The article also cites unnamed “informed sources” saying the model breaks from prevailing VLA, WAM and traditional world-model approaches. It claims the system was trained on only dozens of hours of video data, though the model’s developer has not been identified and the technical claims remain those of the original author and cited source.

A 108-page proof by Harvard mathematician Levent Alpöge and Claude claims to answer one of the longest-running questions in differential and complex geometry: whether the six-dimensional sphere, S⁶, admits a complex structure. The result, described in the source article as a direct construction rather than another attempt along older lines, builds a compact complex threefold X and then identifies it with S⁶ through calculations of the fundamental group, homology arguments, and a chain of classical theorems. The article says mathematician Qiaochu Yuan had GPT-5.6 Sol examine the manuscript. After an initial six-minute pass and another 15 minutes of review, the model reportedly found no flaw. Sol’s view, as cited in the piece, is that if the proof holds up, it may stand as the most important AI-assisted mathematical result so far. The report places this work alongside two earlier Claude-linked advances from July 20 and Aug. 10, and argues that the S⁶ case is different because the geometric object was constructed outright rather than found through search or recombination.

OpenAI has reduced pricing for GPT-5.6 Sol across both API usage and overage credits for the next three months, with the API-side changes already in effect. The biggest cut hit output tokens, which dropped from $30 to $20 per million tokens, while input pricing moved from $5 to $4. Subscription tiers including Pro, Plus, and Business were left unchanged, and included usage was not increased. Cache pricing and long-context pricing were also lowered. The move lands as competition around flagship AI models shifts beyond benchmark performance and into pricing. The original report framed the cut as pressure on Anthropic, which is reportedly in its IPO preparation window, while also pointing to low-cost competition from newer models such as DeepSeek V4 Flash and a model referred to as Ox Alpha. At the same time, OpenAI linked the lower prices to efficiency gains in how Sol now runs. Separately, Codex lead Tibo said Codex weekly active users have passed 20 million. He also said the team has launched a formal investigation into user complaints about fast quota depletion, even though no anomaly has yet been found, and warned that converting subscription accounts into API traffic for resale or shared access is not supported.

Fresh off receiving the Fields Medal, Jacob Tsimerman said at the award ceremony that he plans to join OpenAI to work on AI safety, placing one of mathematics’ highest-profile researchers at the center of a growing debate over what advanced models mean for the discipline. In early August, Tsimerman and OpenAI research scientist Sébastien Bubeck convened roughly 40 mathematicians at OpenAI’s San Francisco headquarters for a closed-door summit, where one talk carried the stark title “The End of Mathematics.” The meeting came after a string of AI-related results in mathematics and computer science. OpenAI said an unreleased model had overturned the unit distance conjecture in May, then disclosed 10 AI-generated findings across several subfields in August. Anthropic, for its part, saw employee Levent Alpöge use Claude to construct orthogonal vector pairs on a 668-dimensional hypercube, while another internal Claude experiment reportedly produced a new result related to the Riemann hypothesis. Those developments have sharpened a divide inside the math community. More than 3,000 mathematicians signed the Leiden Declaration in June, calling for responsible AI use, verification, and proper citation. Others favor rejecting AI tools and AI companies altogether. At issue is not only whether AI can produce proofs, but whether it can explain the hard parts well enough for the work to become part of mathematical literature.

Zhiyuan’s core products — Genie G2, Lingxi X2 and Yuanzheng A3 — took part in the Second World Humanoid Robot Games. In the completed “scenario” events, Zhiyuan’s Genie G2 won two gold medals, three silver medals and two bronze medals. Zhiyuan is currently ranked first on the medal table. The information was reported by Securities Times.

Generalist AI has introduced GEN-1.5, a robotics foundation model the company says can execute new manipulation tasks after watching a single 3- to 12-second demonstration, without additional training or fine-tuning. In tests across 10 short-horizon tasks, the company reported an average one-shot success rate of 59% with a standard deviation of ±10%, and an average few-shot success rate of 83% with a standard deviation of ±9% using roughly five minutes of data, about 50 demonstrations, and 10 gradient steps. The release has drawn comparisons from some researchers to the moment GPT-3 arrived in 2020, though the results remain self-reported and have not been independently verified. The timing also intersects with a broader robotics surge. Unitree went public on Aug. 19 with an opening price of 1,100 yuan per share and a market value of about 444.9 billion yuan, while the World Robot Conference opened in Beijing the same week. Generalist AI, whose backers include Fei-Fei Li as a personal investor and Nvidia as a shareholder, completed a $400 million round in June 2026 at a $2 billion post-money valuation and is reportedly discussing another financing at a $3 billion valuation.

Anthropic and Harvard mathematician Levent Alpöge, working with Claude, have produced a counterexample that the source says overturns the century-old Carathéodory conjecture and, at the same time, the Loewner index conjecture. A PDF that had circulated in mathematics circles for about a week has now been formally released. According to the source, Alpöge and John-Paul Smith constructed a family of functions g_k. In the k=2 case, the support function g₂ defines a convex body with a single umbilic point at the origin whose index is 2. That matters on both fronts: the Loewner conjecture, as described in the source, sets an upper bound of 1 for the index of an isolated umbilic, while the Carathéodory conjecture says a sufficiently smooth closed convex surface in three-dimensional Euclidean space must have at least two umbilic points. The report frames the result as part of a broader shift in mathematical research involving AI systems. It also notes Alpöge’s earlier collaboration with Claude on a claimed counterexample to the 3D Jacobian conjecture and lists other recent AI-related math developments mentioned in the original article.

Anthropic is moving toward an IPO while reshaping its internal control structure. According to The Information, the company plans to issue a class of shares with extra voting rights to CEO Dario Amodei and six other co-founders before listing, marking the first time management would hold super-voting stock in the company’s more than five-year history. The exact voting multiple and issuance size have not been disclosed. The governance move comes as Anthropic posts stronger top-line growth than OpenAI. The article says Anthropic generated $11.6 billion in second-quarter revenue, up from $787 million a year earlier, while OpenAI reported $6.7 billion for the same quarter, versus $5.7 billion in the first quarter. OpenAI’s operating loss reportedly widened from $9.3 billion to $12.3 billion, while Anthropic said its adjusted operating profit turned slightly positive, though the report notes the two profit figures are not directly comparable. The company’s annualized revenue has now crossed $65 billion, and its valuation reached $965 billion after a May financing round. The article says Anthropic confidentially filed an S-1 in June, with Morgan Stanley, Goldman Sachs and JPMorgan as underwriters. Market expectations range from an autumn debut to as early as late September, potentially putting Anthropic on track to list before OpenAI.