Be More with Claude

Gemini 3.5 Flash Review: Blazing fast, but expensive - The speed is honestly insane 550+ tokens/sec and complex React/Tailwind components generate almost instantly - Antigravity 2.0 makes the experience much better with smooth parallel agents and desktop first workflows, although it throttles and randomly shuts down 1-2 subagents - It feels strong for frontend work, quick UI generation, But once you move into harder architecture and reasoning heavy tasks, the weaknesses start showing - Hallucinations happen more often on complex problems, and Gemini 3.1 Pro still feels smarter for deep coding, system design, and long reasoning workflows - On benchmarks, Gemini 3.5 Flash reached #9 on the Text + Code Arena frontend leaderboard, but honestly at this cost I expected much better - The biggest issue is pricing benchmark costs are estimated around $1,552 while GPT-5.5 Medium is cheaper, and any day I would choose GPT-5.5 over Gemini 3.5 Flash - But my suggestion right now is Composer 2.5, it looks like the better value overall and gets surprisingly close to Opus 4.7 Max and GPT-5.5 xhigh quality while being almost 20x cheaper

Gemini 3.5 Pro leaks: Google's frontier model is almost here - It officially confirmed Gemini 3.5 Pro is coming next month, likely around mid-to-late June - It will become Google's new flagship SOTA model focused heavily on coding, agents, reasoning, and memory - Pricing will stay close to Gemini 3.1 Pro or around 1.5x higher - It won't be a Mythos-style model, but more in the GPT-5.5 / Claude Opus 4.7 - It also heavily focused on grounding + search reliability to reduce hallucinations in real world usage - Expecting faster responses than Gemini 3.1 Pro with much stronger long-horizon agentic workflows - Hoping this time the knowledge cutoff is newer than just January 2026

Six months ago: $150K and 8 weeks to deploy an enterprise voice AI agent. Today: 10 minutes, free, from a website URL. Quick thread on what just changed and why it matters 🧵 https://t.co/mIG0gzQ9Mc