Eight AI Updates: Math Claims, Model Pricing, Fake CVEs, Robots, and Compute

<p>AI news on August 3 ranged from ambitious mathematics claims to a warning about fabricated vulnerability reports entering formal data systems. Here are eight developments, with the source boundary attached to each one.</p> <h3> 1. OpenAI published ten mathematics and theoretical computer science advances </h3> <p>OpenAI says an internal version called Astra solved or materially advanced ten long-standing open problems. According to the company, the model generated mathematical arguments, humans helped organize them into papers, and the model then formalized the work into Lean certificates. OpenAI estimated that the tokens used to find the solutions would cost about $2,000 at Sol API prices.</p> <p>This remains an OpenAI publication, not a settled academic verdict. Correctness, novelty, and the status of each result still require review by the relevant research communities.</p> <p>Source: <a href="https://openai.com/index/ten-advances-in-mathematics/" rel="noopener noreferrer">https://openai.com/index/ten-advances-in-mathematics/</a></p> <h3> 2. Qwen3.8-Max was unveiled, but its open weights are not available yet </h3> <p>Alibaba announced Qwen3.8-Max with 2.4 trillion total parameters and 95 billion active parameters. The published specification includes text, image, and video input plus a one-million-token context window. The weights are expected next week.</p> <p>Those specifications and performance claims are vendor statements. The release of the weights has not happened yet, and benchmark or long-running task results should not be described as independently verified.</p> <p>Source: <a href="https://qwen.ai/blog?id=qwen3.8" rel="noopener noreferrer">https://qwen.ai/blog?id=qwen3.8</a></p> <h3> 3. A third-party test put DeepSeek V4-Flash at a very low token price </h3> <p>Reuters cited Artificial Analysis measurements that priced V4-Flash at about $0.14 per million input tokens and $0.28 per million output tokens. The reported average cost was about $0.03 per benchmark task.</p> <p>This is a third-party benchmark and pricing snapshot. Token prices alone do not establish the total cost, reliability, or value of a model on a real workload.</p> <p>Source: Reuters, August 3, 2026, reporting Artificial Analysis measurements. The source brief did not include the article URL.</p> <h3> 4. JFrog says 54 reviewed advisories appeared to be LLM fabrications </h3> <p>JFrog investigated a new GitHub repository that published more than 50 vulnerability advisories. The researchers concluded that 54 advisories appeared to be LLM-generated fabrications. The reports cited missing functions, unrelated code lines, or proof-of-concept inputs that failed before reaching the alleged vulnerable code. Several reports still entered downstream systems and received high or critical severity metadata.</p> <p>The finding applies to this batch. It does not show that the entire CVE system is untrustworthy. JFrog's strongest evidence came from checking source code and PoCs, not from relying on an AI detector.</p> <p>Source: <a href="https://research.jfrog.com/post/sqlite-critical-cves-or-llm-slops/" rel="noopener noreferrer">https://research.jfrog.com/post/sqlite-critical-cves-or-llm-slops/</a></p> <h3> 5. Google released Gemini Robotics ER 2 for high-level robot reasoning </h3> <p>Gemini Robotics ER 2 is available through the Gemini API and Google AI Studio. Google says it supports continuous video progress tracking, self-correction, tool use, and coordination among multiple robots. Its role is high-level reasoning rather than direct motor control.</p> <p>Google's enterprise platform remains in private preview. Public demonstrations and evaluations are company-led, and real performance depends on the robot hardware and control stack.</p> <p>Source: <a href="https://blog.google/innovation-and-ai/models-and-research/google-deepmind/gemini-robotics-er-2/" rel="noopener noreferrer">https://blog.goog