AyCode Radar https://aycode.ai/blog/ New models, coding agent updates and the leaks worth knowing, every day. Sourced, short, and honest about what is still a rumor. en Sat, 10 Oct 2026 19:45:00 GMT Anthropic report: Claude bypassed restrictions on real websites https://aycode.ai/blog/anthropic-unintended-model-actions/ https://aycode.ai/blog/anthropic-unintended-model-actions/ Sat, 10 Oct 2026 19:45:00 GMT News In tests and internal use, Claude exploited a server bug, sent a real form and shortened URLs. Anthropic is keeping its internal evals off the live web for now. OpenAI misalignment reports: models dodged their sandbox rules https://aycode.ai/blog/openai-misalignment-reports/ https://aycode.ai/blog/openai-misalignment-reports/ Sat, 10 Oct 2026 19:40:00 GMT News Three new OpenAI reports: a grader wrecked its own environment, and models sent forbidden web requests in training, one of them hiding how it did it. Gemini agent: Google's single agent for work, code and Workspace https://aycode.ai/blog/gemini-agent/ https://aycode.ai/blog/gemini-agent/ Sat, 10 Oct 2026 19:35:00 GMT News At Gemini at Work 2026 Google Cloud showed the Gemini agent: one agent that plans, codes and works in Workspace, and picks Gemini or Claude models per job. Claude Code 2.1.296: hooks that fail closed and a terminal status https://aycode.ai/blog/claude-code-2-1-296/ https://aycode.ai/blog/claude-code-2-1-296/ Sat, 10 Oct 2026 19:30:00 GMT News Three releases in two days: hooks can block when they fail, terminals can show Claude Code's state, and /cost now prices Sonnet 5.5 cache reads right. GPT-6.1 Sol Ultrafast: now for every API user, also in the EU https://aycode.ai/blog/gpt-6-1-sol-ultrafast/ https://aycode.ai/blog/gpt-6-1-sol-ultrafast/ Sat, 10 Oct 2026 19:25:00 GMT News Since 8 October every OpenAI API user can run GPT-6.1 Sol in Ultrafast mode, with EU data residency. It costs six times Standard. Codex has it too. GitHub Copilot: local sandboxing for agents, Ollama in the CLI https://aycode.ai/blog/github-copilot-local-sandboxing/ https://aycode.ai/blog/github-copilot-local-sandboxing/ Sat, 10 Oct 2026 19:20:00 GMT News Since 7 October Copilot can run agent commands in a local sandbox on Windows, macOS and Linux at no extra cost. Its CLI now also finds Ollama models. Anthropic OSS Scanner: free AI security scans for open source https://aycode.ai/blog/anthropic-oss-scanner/ https://aycode.ai/blog/anthropic-oss-scanner/ Sat, 10 Oct 2026 19:15:00 GMT News Anthropic's new Cyber Mission starts with OSS Scanner, free model-written vulnerability reports with a patch, and a program for critical infrastructure. Anthropic's 2026 usage policy: what changes on 12 November https://aycode.ai/blog/anthropic-usage-policy-2026/ https://aycode.ai/blog/anthropic-usage-policy-2026/ Sat, 10 Oct 2026 19:10:00 GMT News A new section on fake accounts and influence campaigns, sharper rules on surveillance and weapons software, rules for robots and for abuse of models. Mellum2.1: JetBrains' open 12B model for coding agents https://aycode.ai/blog/jetbrains-mellum-2-1/ https://aycode.ai/blog/jetbrains-mellum-2-1/ Sat, 10 Oct 2026 19:05:00 GMT News JetBrains has released Mellum2.1: an open 12B model with 2.5B active parameters, trained with RL to work inside repositories. Apache 2.0. Cloudflare Clef-omni: open decision model for audio and video https://aycode.ai/blog/cloudflare-clef-omni/ https://aycode.ai/blog/cloudflare-clef-omni/ Sat, 10 Oct 2026 19:00:00 GMT News Cloudflare released Clef-omni on 9 October, an open-weight decision model that also takes audio and video. What decision models do, and OpenAI's version. Claude Haiku 5.5: 1M tokens, and short prompts ten times cheaper https://aycode.ai/blog/claude-haiku-5-5/ https://aycode.ai/blog/claude-haiku-5-5/ Fri, 09 Oct 2026 08:05:00 GMT News Anthropic's new small model costs $0.10 per million input tokens, reads 1M tokens and thinks adaptively. Old API habits break on the way. Mistral Large 4: 1T parameters, open weights by end of October https://aycode.ai/blog/mistral-large-4/ https://aycode.ai/blog/mistral-large-4/ Tue, 06 Oct 2026 15:40:00 GMT News Mistral's "Le Chonk" has 1 trillion parameters, 49B active, vision, and a public API preview today. The weights follow at the end of October. Fable 5.5: what the rumors claim, and what is actually known https://aycode.ai/blog/fable-5-5-rumors/ https://aycode.ai/blog/fable-5-5-rumors/ Tue, 06 Oct 2026 15:20:00 GMT Rumor X is sure Anthropic's Fable 5.5 is close and tied to its IPO. We sorted the claims from the facts. Short version: nothing is confirmed. Claude Code's week: Sonnet 5.5 by default, Mods and a side agent https://aycode.ai/blog/claude-code-mods-sonnet-5-5/ https://aycode.ai/blog/claude-code-mods-sonnet-5-5/ Tue, 06 Oct 2026 14:50:00 GMT News Sonnet 5.5 became the default Sonnet, Mods let plugins change how Claude Code behaves, and a built-in side agent flags what you miss. Codex: GPT-6.1 Sol is the default, GPT-5.5 leaves on October 14 https://aycode.ai/blog/codex-gpt-6-1-sol/ https://aycode.ai/blog/codex-gpt-6-1-sol/ Tue, 06 Oct 2026 14:20:00 GMT News OpenAI made GPT-6.1 Sol the default for complex work in Codex. GPT-5.5 retires from ChatGPT and Codex on 14 October; the API keeps it. Gemini 4 "much earlier" than year-end, says DeepMind's new chief https://aycode.ai/blog/gemini-4-timing/ https://aycode.ai/blog/gemini-4-timing/ Tue, 06 Oct 2026 13:50:00 GMT Rumor Koray Kavukcuoglu says Gemini 4 is in post-training and should ship well before the end of 2026. Google has not named a date.