Anthropic's Claude Haiku 5.5 Cuts Costs 75%, Betting the Agent Era Belongs to Small Models
Anthropic shipped its fastest, cheapest small model yet — built explicitly for subagent swarms and browser use — and paired it with an official multi-agent architecture that reads like a blueprint for how the next generation of AI systems should be wired.
Anthropic released Claude Haiku 5.5 on Wednesday, positioning it not as a flagship but as infrastructure. The model costs roughly 75% less to run than its predecessor, Haiku 4.5, according to @ChrisPirillo, who noted it was "built for subagents and browser use." Anthropic's own framing, surfaced by @CrazyAIAgent, calls it "the cheapest, fastest, and most capable small model we've ever released." The emphasis is deliberate. This is a release about economics, not leaderboards.
The pricing math matters more than it might appear. In a multi-agent system, a single user request can fan out into dozens or hundreds of model calls as subagents parallelize work — reading files, browsing pages, drafting candidate answers. At Haiku 4.5 prices, that fan-out was a real constraint on how aggressively builders could architect swarms. A 75% cut changes the calculus. Suddenly it's economically rational to throw ten cheap agents at a problem where you previously would have used one expensive one.
Get our free daily newsletter
Get this article free — plus the lead story every day — delivered to your inbox.
Want every article and the full archive? Upgrade anytime.
No spam. Unsubscribe anytime.