Anthropic Unleashes Claude 3.5 Upgrades with a Dash of AI Wizardry

Today marks the release of Anthropic’s Claude 3.5 Sonnet, now boasting coding prowess that leaps from 33.4% to 49.0% on SWE-bench Verified. Enter Claude 3.5 Haiku, matching the former Claude 3 Opus performance without breaking the bank. The pièce de résistance? The experimental 'computer use' feature in public beta, letting AI navigate screens like a tech-savvy human. Early adopters praise its potential, even as Anthropic ensures safety with US and UK AI Safety Institutes. Available via major APIs and free on Claude.ai, it’s AI evolution with a wink and a nod.