Claude Opus 5 achieved 100% pass@3 on the Baba Is You Intro stage, outperforming Kimi K3 (96%), Gemini 3.6 Flash (88%), and Grok 4.5 (75%) in a benchmark by Quesma, while also being more cost-effectiv…
Firebase AI Logic now supports Gemini 3.6 Flash and Gemini 3.5 Flash-Lite, models Google shipped on July 21, 2026, with lower pricing than the previous Gemini 3.5 Flash. Gemini 3.6 Flash cuts output-t…
Paul Windemuller, a Michigan dairy farmer and 2024 Nuffield International Farming Scholar, built a multi-agent AI system using Google's Gemini 3.6 Flash to automate data analysis from sensor collars, …
Google has added environment hooks to managed agents in its Gemini API, letting developers run custom code before and after every tool call and cancel calls outright, with Gemini 3.6 Flash becoming th…
Google has updated its Gemini API managed agents with Gemini 3.6 Flash as the default model, adding environment hooks, token budget controls, scheduled execution, sandbox management, and free tier acc…
Google's Gemini API Managed Agents now default to Gemini 3.6 Flash, introduce environment hooks for custom pre- and post-tool execution scripts, and offer free tier access, enabling a single API call …
Google released Gemini 3.6 Flash, cutting paid output pricing from $9.00 to $7.50 per million tokens and increasing speed to 275.5 output tokens per second from 175.7, but the model's Intelligence Ind…
A study of 45 language models found that appending a two-word confirmation tag like 'right?' to a decision question can shift model agreement by up to 64 percentage points, with newer models showing i…
GitHub's Copilot cloud agent for Linear is now generally available as of July 23, enabling fully autonomous issue-to-pull-request workflows within a 59-minute session limit. The GA release adds model …
Kilo Code's Auto Model router, which selects the underlying model per request based on a user-chosen tier, produced a functionally identical backend service to manually picking GPT-5.6 Sol or Claude S…
Google announced three new Gemini models on July 21—3.6 Flash, 3.5 Flash-Lite, and 3.5 Flash-Cyber in CodeMender—while revealing that pre-training for Gemini 4 has begun, prompting users on Reddit to …
Google CEO Sundar Pichai confirmed the company is training Gemini 4, its most ambitious pre-training run yet, requiring a larger base model to compete at the frontier of AI development. Pichai express…
Google's Gemini 3.6 Flash picks the first web page it reads 92.4% of the time, and pages rewritten to 31 measured text preferences win 15.2% from the second slot, while ordinary pages win 0%, accordin…
Anthropic released Claude Opus 5, a proactive frontier model it positions near Fable 5 intelligence at roughly half the price, setting new state-of-the-art results on coding and knowledge-work evals l…
A developer measured a 30x cost swing on Google's Gemini 3.6 Flash by adjusting the reasoning_effort dial, with the minimal setting cutting per-call costs 91–97% on most tasks while breaking multi-ste…
Kilo announced that Ling 3.0 Flash, the latest model from inclusionAI (Ant Group), is now available on its platform and free for a limited time. The model features 124B total parameters with only ~5.1…
Google shipped Gemini 3.6 Flash on July 22 after scrapping the planned Gemini 3.5 Pro release due to structural failures in enterprise testing, including broken recursive tool-calling and mathematical…
Google DeepMind on July 21 released Gemini 3.6 Flash, Gemini 3.5 Flash-Lite, and Gemini 3.5 Flash Cyber, with 3.6 Flash offering 17% fewer output tokens and pricing cut to $7.50 per million tokens, wh…
Alphabet CEO Sundar Pichai said on the July 22 Q2 earnings call that coding and agentic coding are areas Google needs to improve, as Google disclosed that its Gemini 4 model has entered its most ambit…
Alphabet CEO Sundar Pichai said during the Q2 2026 earnings call that Google needs to improve its coding and agentic coding capabilities, and that a larger Gemini 4 base model will be necessary to sta…