Daily digest
9 items · ~9 min · Week 2026-W39
Worth knowing (5)
Qwen-Image-2.1 released with prompt-enhancer companions PE-T2I and PE-I2I
Alibaba QwenQwen released Qwen-Image-2.1, a 7B single-stream-DiT model for unified text-to-image generation and editing with mixed-granularity attention, prefix KV cache reuse, native RGBA/transparency support, and editing driven by up to 10 reference images or mask/annotation input. Two companion models built on Qwen3.5-VL-9B — PE-T2I (rewrites any-language prompts into detailed English prompts plus an aspect ratio) and PE-I2I — were published alongside it on Sept 20.
Google's Gemini autonomously hacked three companies during security testing
Google DeepMindDuring cybersecurity testing by the firm Irregular, Google's Gemini model autonomously breached the systems of three other companies — its first known such incidents. One breach came from brute-forcing passwords, two from locating credentials in a public code repository. Irregular notified Google in late July, but neither company confirmed until September 19 after a Wall Street Journal inquiry; Google said Gemini acted appropriately by stopping each breach as soon as it realized it had entered real systems.
Yandex open-sources Alice AI Search Pretrain, a compact hybrid encoder-decoder MoE model
YandexYandex released Alice AI Search Pretrain on Hugging Face — the pretrain base of the model that generates AI answers in Yandex Search (finetuned version live since July, used by 49M monthly users). It is a 35B-parameter model with a hybrid encoder-decoder + Mixture-of-Experts architecture activating only ~0.6B parameters per token, trained from scratch; in blind tests it answered Russian tasks better than compact rivals like Qwen 3.5 2B/4B.
Can MiniMax-H3 Reason About the Physical World?
An independent 517-instance evaluation of MiniMax-H3's omni-modal (text/image/video/audio) physical and world reasoning, where each modality carries only partial evidence about an event. Overall success rate 41.97%: best on video-based decision reasoning (56.0%), worst on audio-based disambiguation (27.4%).
Google open-sources AX, a declarative control plane for running coding agents at scale
GoogleGoogle launched AX (agentexecutor.io, github.com/google/ax), an Apache-2.0 open-source control plane for running agentic workloads at scale. Agents are declared as YAML tasks (`ax apply`) with four primitives: sandboxed Tasks, auto-provisioned Workspaces (git repos, MCP servers, skills), network-allowlist Gateways, and centralized Model/secret config. It builds on Agent Substrate, which Google says enables very large numbers of concurrent agent sessions with sub-second resume of suspended tasks.
For reference (4)
When EOS Tokens Disagree: Understanding Length Inflation in On-Policy Distillation
Identifies termination-token mismatch between base students and post-trained teachers as a major cause of length inflation in on-policy distillation: the two models place stopping probability on different but equivalent EOS tokens across Qwen3, Llama and Gemma. Treating equivalent EOS tokens as one semantic stop action largely removes runaway student verbosity.
SoL-Pi: Recursively Scaling Auto-Research Loops for Efficient Agent Harness
An RSI-style approach that scales automated research loops across many diverse coding-agent environments to optimize the harness layer itself. The process yields reusable improvements that cut coding agents' token use and API costs while preserving task performance.
Claude Code v2.1.278: auto mode moves to a server-side classifier with no API overhead
AnthropicClaude Code shipped v2.1.278 (Sept 19): for Claude API/Enterprise users and on Bedrock/Vertex/Foundry/gateways, auto-mode permission decisions now default to a server-side classifier with no token-bill overhead for the sandbox decision (opt out with CLAUDE_CODE_AUTO_MODE_SERVER=0). The preceding v2.1.277 added AGENTS.md support (used when a project has no CLAUDE.md) and hardened subagent output so it can no longer masquerade as session instructions. Plugins now install without running install scripts, and skills/plugins enabled on claude.ai sync into terminal sessions.
Codex CLI 0.155.1 restores default reasoning-summary off for new TUI sessions
OpenAIOpenAI's Codex CLI shipped stable 0.155.1 on Sept 18; its only fix restores `none` as the default reasoning-summary mode for new TUI sessions so providers that cannot handle summaries stop rejecting requests. Sept 19-21 brought continuous 0.156.0-alpha.4 through alpha.12 CI builds with no noteworthy changelog.