Anthropic Launches Claude Haiku 5.5, Cutting The Cost Of Its Smallest Model By Around 75%: All Details
Anthropic has released Claude Haiku 5.5, which it calls its cheapest, fastest and most capable small model to date. The company says the model costs around 75 percent less to run, on average, than Haiku 4.5, and is available now on all platforms, including Amazon Web Services, Google Cloud and Microsoft Azure. The launch is as much about economics as capability. Small models handle the bulk of…
Anthropic has unveiled Claude Haiku 5.5, heralded as the company's most affordable, quickest, and capable small model yet. The model will cut costs by around 75 percent on average compared to its predecessor, Haiku 4.5, marking a significant shift in accessibility for businesses. Available across major platforms, including Amazon Web Services, Google Cloud, and Microsoft Azure, the release underscores cost-effectiveness as a key factor in AI adoption.
Pricing for Haiku 5.5 has been recalibrated, with input tokens at $0.10 per million and output tokens at $0.50 per million, covering the majority of requests. For prompts exceeding 100,000 tokens, the pricing drops to 50 percent and 90 percent of earlier model costs, respectively. Anthropic explains that an enhanced tokenizer accounts for a slightly higher token usage per task, which is already factored into the average 75 percent cost reduction.
Targeted at high-volume, cost-sensitive tasks like summarization, data processing, and classification, Haiku 5.5 also serves as a subagent for coding alongside Opus 5.5 and Sonnet 5.5. The model introduces an adjustable effort setting, enabling users to balance cost and intelligence. Haiku 5.5 has demonstrated impressive performance, achieving scores of 72.4 percent on OSWorld 2.1 and 39.2 percent on Terminal-Bench 4.0, showcasing superior efficiency compared to its predecessor and competitors.
Anthropic has bolstered its cybersecurity measures, maintaining a balance between defensive capabilities and accessibility, while also offering a $100 million commitment to train 10,000 AI engineers through a new residency program. Early adopters have reported notable speed gains, such as a 30 percent lower latency and up to 2.5 times faster inference per agent turn with Asana.
To further incentivize developers, Max and Team subscribers will receive monthly API credits, with varying tiers based on subscription levels. The Python and TypeScript SDKs include beta support for advanced use cases, making Claude Haiku 5.5 accessible via `claude-haiku-5-5`.
Written by urgent.news from Free Press Journal's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.