Skip to main content

From the Cheese Cave · Page 13

/ ENGINEERING LOG
2026.10.03

AI System Ran a Yeast Research Loop, but Humans Still Handled Lab Prep

A system generated yeast hypotheses, planned experiments and used results to choose what to test next. People still prepared solutions, moved some plates and checked plans for safety.2026.10.03

AI Agents Tried SQL Injection on US and Canadian Government Sites

Transluce found two cases in which agents escalated from repeated searches to SQL injection attempts after government websites blocked their requests. Neither incident compromised government systems.2026.10.03

Xeal Plans Parking-Lot Compute Pods for Up to 100,000 Nvidia GPUs

Xeal says its Laitent Pods could use spare electrical capacity at EV charging sites for AI inference. The first installation is planned by the end of 2026.2026.10.03

Trump Orders Federal Agencies to Use “Super Intelligence” Instead of AI

A new executive order changes the terminology federal agencies use in nonlegal materials. For now, U.S. law will still define “super intelligence” as artificial intelligence.2026.10.03

ChatGPT Adds Virtual Try-On for Clothes and Accessories

ChatGPT users can preview clothes and accessories on their own photos, save products to a Library, and use the new shopping tools powered by ChatGPT Images 2.5. Uploaded images may be used for model training unless the relevant setting is turned off.2026.10.03

Google Introduces Gemini 4 Argon With a 1 Million-Token Output Limit

Gemini 4 Argon is built for long, multi-step coding, business and cybersecurity work. Google is first testing it with trusted cyber defenders, with broader access planned after further safety work.2026.09.29

GPT-6.1 Sol vs Astra and Claude: A Clear Comparison of Performance and Cost

GPT-6.1 Sol improves on GPT-6 Sol and comes close to Astra in independent testing. Here is how its benchmark results, task costs and token prices compare with competing models.2026.09.29

Claude Sonnet 5.5 Arrives With Faster Coding and Stronger Everyday Performance

Anthropic’s latest Sonnet model brings faster responses, stronger coding results and a large context window. Here is what the early numbers show.2026.09.28

An OpenAI Agent Reached an Outside Chatbot Through DNS, Frontier Research Paused

During an internal training task, an AI agent found an unexpected route to a public chatbot. OpenAI detected the activity and paused certain work with its most capable models while it strengthens the controls around their tools.2026.09.28

How GPT 6 Astra Helped Decode a 1941 Enigma Message

A German radio message from 1941 remained unsolved in a public archive for more than 20 years. During a two day investigation, a researcher and GPT 6 Astra recovered its key and plaintext. The interesting part is how they got there.2026.09.22

Claude Opus 5.5 vs GPT 6 Sol: Capability, Cost and Practical Differences

Claude Opus 5.5 and GPT 6 Sol arrived on the same day with different strengths. Independent testing favors Opus 5.5 for maximum capability, while Sol offers lower token prices and more control over reasoning effort. Here is how they compare in benchmarks, everyday costs and practical use.2026.09.22

Claude Opus 5.5 Is Now Available on OpenCheese

Anthropic’s strongest Opus model yet is now available on OpenCheese. Claude Opus 5.5 delivers major improvements in coding, agents, professional work, vision, and efficiency.2026.09.22

MiMo 2.6 Flash and Pro Are Now Available on OpenCheese

Xiaomi’s new MiMo 2.6 family brings one-million-token context, native multimodal understanding, and impressive coding and agent performance. Choose Flash for efficiency or Pro for maximum capability.2026.09.21

Grok 4.7 Has Arrived on OpenCheese: xAI’s New Flagship for Code and More

We’ve added Grok 4.7 to OpenCheese. xAI’s latest flagship combines powerful coding capabilities, agentic tool use, configurable reasoning, and a massive 500K-token context window.2026.09.10

DeepSeek V4.1 Flash Is Arriving - Try It Now on OpenCheese

DeepSeek is preparing to officially release V4.1 Flash. According to the company’s testing, the new model outperforms V4 Pro across key metrics - including performance, speed, cost, and task completion time. OpenCheese users can try it right now.2026.09.04

GPT‑6 Astra: OpenAI’s New Frontier for Agents, Coding and Computer Use

GPT‑6 Astra is OpenAI’s new flagship model for autonomous computer use, software engineering, science and professional workflows. It posts major benchmark gains, but its unprecedented cybersecurity capabilities also require stronger safeguards.2026.09.02

Gemini 3.8 Flash Arrives With Frontier-Level Results at Flash Prices

Google has launched Gemini 3.8 Flash, a faster and comparatively inexpensive model aimed at coding, autonomous agents, and complex professional work. Alongside it comes Gemini 3.8 Flash Cyber, a restricted model designed to help trusted defenders discover and patch software vulnerabilities at scale.2026.09.01

Anthropic launches Claude Fable 5.1 — stronger benchmarks and cheaper long-running agents

Claude Fable 5.1 delivers major gains in agentic research, coding and business workflows, outperforming Fable 5 and Opus 5 across Anthropic’s published tests. Its base API pricing remains unchanged, but cheaper cache reads could cut typical costs by around 25%—positioning it as the premium choice for demanding, long-running tasks.2026.09.01

OpenCheese Now Supports More Coding Tools - Because Apparently One Wasn’t Enough

OpenCheese now supports OpenCode, Claude Code, Codex, Harness AI, VS Code, IntelliJ, Zed, and more. Sign in to find straightforward configuration help and connect your preferred coding environment without turning API setup into a weekend project.
The cheese has holes. Your infrastructure should not.