Meta released Muse Spark 1.1, a frontier model priced at $1.25 per million input tokens and $4.25 per million output tokens, undercutting rivals like GPT-5.5 and Claude Opus 4.8 by over 80%. The launc…
Researchers propose Graph-Regularized Agentic Context Evolution (GRACE), a method that maintains persistent LLM agent instructions as a typed semantic graph to enable local verification of updates, im…
Researchers at ASSET Research Group published a proof-of-concept attack called GhostCommit that exploits coding-agent pipelines by hiding exfiltration instructions inside image files in pull requests.…
Researchers evaluated Gemini models as audio judges for full-duplex voice agents, finding that Gemini 2.5 Flash achieves near-human reliability on most scoring dimensions. The study shows that LALM (L…
Meta Superintelligence Labs released Muse Spark 1.1, a multimodal reasoning model for agentic tasks, alongside a public preview of the Meta Model API. The closed, hosted model offers adjustable reason…
OpenAI released GPT-5.6 Sol, alongside Terra and Luna, ending a two-week preview restricted by the U.S. Department of Commerce. Sol in ultra mode tops Terminal-Bench 2.1 at 91.9% and matches Anthropic…
A developer tested ChatGPT (GPT-5.2), Claude (Opus 4.8), Gemini (3.1 Pro), and Grok 4 on the same real coding tasks and found that the best model depends on the specific job. For debugging, reasoning-…
Google Developer Expert Andriy Lukyanyuk tested Google Antigravity 2.0 with Gemini 3.1 Pro by entering a Kaggle Playground Series competition. The agentic IDE built a pipeline with LightGBM, CatBoost,…
Researchers released TaxCalcBench, an open-source benchmark evaluating AI models' ability to file taxes using realistic PDF inputs for federal and state returns. GPT-5.4 Pro achieved the highest stric…
Google updated its Android Bench benchmark for AI agents in Android development, adding eight new models including Claude Fable 5, which leads with 84.5% accuracy. The update introduces cost and effic…
Alibaba's Qwen team released Qwen-AgentWorld, a family of language world models trained to simulate agentic environments by predicting next states as text, outperforming GPT-5.4 and Claude Opus 4.8 on…
Anthropic released Claude Opus 4.7 on April 16, 2026, a targeted upgrade focused on making agentic coding loops and multi-step tool use production-ready. The model achieves 87.6% on SWE-bench Verified…
Meituan released LongCat-2.0, a 1.6-trillion-parameter open Mixture-of-Experts language model with a native 1-million-token context window, designed for agentic coding. The model activates about 48 bi…
GitHub will deprecate Gemini 2.5 Pro and Gemini 3 Flash across all Copilot experiences on July 31, 2026, recommending Gemini 3.1 Pro and Gemini 3.5 Flash as alternatives. Users must update workflows a…
Anthropic launched Claude Sonnet 5, its most agentic mid-tier model, now available in Kilo Code via OpenRouter and direct API. The model achieves 63.2% on SWE-bench Pro and edges past Opus 4.8 on know…
Anthropic released Claude Sonnet 5, a midsize AI model with advanced autonomous capabilities, priced at $2 per million input tokens through Aug. 31. The model achieves 80.5% on agentic coding benchmar…
OpenAI released GeneBench, a benchmark of 103 multi-stage genomics problems, revealing that top AI models like GPT-5.5 Pro achieve only a 33.2% pass rate, with 60% of problems below 20% for current mo…
Meituan launched LongCat-2.0, a 1.6-trillion-parameter AI model trained entirely on domestically developed chips, claiming it as the first of its scale to achieve end-to-end training on Chinese silico…
A study evaluating AI tools on real clinical queries from physicians found that a specialized clinical tool (OpenEvidence) outperformed frontier general-purpose models (Claude Opus 4.8, Gemini 3.1 Pro…
Google DeepMind released Gemini 3.1 Pro on February 19, 2026, enabling note-taking in Google Meet for Google AI Pro and Ultra subscribers. The model ranks highly on benchmarks, including #2 in reasoni…