Anthropic Launches Claude Sonnet 5: Near-Opus Performance at One-Third the Price

Anthropic says Sonnet 5 is nearly as powerful as its flagship Opus 4.8 — at a third of the cost.

Three Key Takeaways

Anthropic officially released Claude Sonnet 5 on June 30, making it the default model for all Free and Pro users starting July 1. The model shows significant improvements over Sonnet 4.6 in reasoning, tool use, coding, and knowledge work. Anthropic calls it "the most agentic Sonnet ever" — designed not just to answer questions but to autonomously plan, use browsers and terminals, and execute multi-step tasks.

The pricing is aggressively competitive. The introductory rate through August 31 is $2 per million input tokens and $10 per million output tokens, rising to $3/$15 afterward. Compare that to Opus 4.8 at $15/$75 — Sonnet 5 delivers close performance on most tasks at less than one-fifth the cost. It's available to Max, Team, and Enterprise customers as well.

Safety systems got a simultaneous upgrade. Sonnet 5 ships with real-time cybersecurity safeguards enabled by default, designed to detect and block dangerous use cases. Anthropic frames this as a key piece of its "responsible AI" narrative ahead of a potential blockbuster IPO.

WangDou's Take

The real message of this launch isn't "Sonnet 5 is strong." It's "Is Opus 4.8 worth the premium?" When a mid-tier model handles 80–90% of what the flagship does at one-fifth the price, the flagship's value proposition shrinks to "How much will you pay for that last 10%?" Anthropic knows this — the limited-time discount is a classic lock-in play: flood the user base, get developers to hardwire their code against Sonnet 5's API, and by the time prices normalize in September, switching costs make inertia the default. It's the same logic as a carrier phone contract, except what's locked in isn't a SIM card — it's your system prompts and tool schemas.

Source: Anthropic Blog · VentureBeat