Focused on sharing: AI programming | Agents | Tools
Anthropic’s limited‑preview Claude Code Security, an AI agent that reads and patches code, triggered a sharp sell‑off in major cybersecurity stocks, while its ability to uncover hundreds of hidden bugs raises questions about the future role of traditional security firms and junior analysts.
Anthropic unveiled Claude Sonnet 4.6 on February 18, touting Opus-level coding ability, a 1 million-token context window, and unchanged pricing; benchmarks show a SWE-bench score of 79.6% (up from 77.2%), OSWorld 72.5% (vs 61.4%), and GPQA Diamond 89.9%, while industry leaders praise its reduced laziness, stronger instruction following, and strategic long-term planning.
On Chinese New Year's Eve, Alibaba's Qwen 3.5 open‑source model—featuring a 397 billion‑parameter backbone with a 17 billion‑parameter active set, hybrid linear attention, and sparse MoE—was released under Apache 2.0, delivering 8.6‑19× faster inference, top‑tier agent, code and multimodal scores, and rapid integration across major AI platforms.
A leaked SWE‑Bench score of 83.7% for DeepSeek V4 sparked claims it outperforms Claude Opus 4.5 and GPT‑5.2, but the data was later debunked as fabricated while official hints confirm a 1‑million‑token context model and a mid‑February 2026 release.
Peter Steinberger, the creator of the fast‑growing open‑source AI agent project OpenClaw, announced his move to OpenAI, a shift that many view as a major strategic error by Anthropic after it sent legal letters demanding a rename, while OpenAI pledged to keep the project open and independent.
The MiniMax M2.5 model, released open‑source on Feb 13, achieves an 80.2% SWE‑Bench Verified score that surpasses GPT‑5.2, Claude Opus 4.6 and Google Gemini 3 Pro, runs 37% faster than its predecessor, costs only $1 per hour, and demonstrates SOTA agent abilities in browsing and tool use, marking a major leap for Chinese large‑language models.
ByteDance's Seed 2.0 Pro (Doubao 2.0) showcases industry‑leading performance on math, vision, document, long‑video, and code benchmarks, dramatically lowers inference cost, and is now available in the Doubao app and Trae IDE, positioning it as a serious challenger to GPT‑5.2 and other top LLMs.
After Claude Code was updated to version 2.1.37, developers observed a sharp drop in context‑caching hit rates, causing unexpected cost spikes and slower responses, and a community investigation revealed that random headers and spaces injected by the tool break the model’s cache matching.
Anthropic announced a $30 billion G‑round that lifts its valuation to $380 billion and its run‑rate revenue to $14 billion, while Elon Musk blasted the company on X for allegedly embedding anti‑Chinese, anti‑white, and anti‑male bias in its AI models, sparking a wider debate over woke culture and AI alignment.
Z.ai released the open‑source GLM‑5 model with 744 billion parameters, 28.5 T tokens of training data, and new Sparse Attention and Slime RL infrastructure, achieving top open‑source rankings and near‑Claude Opus 4.5 performance on Vending Bench 2 and CC‑Bench‑V2 while adding multi‑scenario agent capabilities.
