Claude Haiku 5.5
Claude Haiku 5.5 is Anthropic's fastest small model for high-volume tasks and subagent work, with adjustable effort, adaptive thinking, and 1M context.
Model Overview
Capabilities, design details, and architectural traits
Claude Haiku 5.5 - Anthropic's fastest small model for high-volume subagent work
Claude Haiku 5.5 is the fastest and most capable small model Anthropic has released. It is built for high-volume workloads: summaries, compactions, database queries, classification, and routing. Its speed makes it a fit for latency-sensitive uses like live customer support and browser use. Anthropic positions it as the subagent layer of the Claude 5.5 family, pairing with Opus 5.5 and Sonnet 5.5 on coding work, where the larger model plans and Haiku 5.5 executes well-defined tasks quickly and in parallel.
What sets it apart
| Trait | Detail |
|---|---|
| Fastest Claude model | Anthropic's fastest model to date at standard speed, ahead of Sonnet 5.5 and Opus 5.5 |
| First Haiku with effort controls | Adjustable effort setting lets users optimize each task, a first for the Haiku class |
| Subagent pairing with Opus 5.5 | Documented workflow where Opus 5.5 plans and makes judgment calls while Haiku 5.5 runs quick, high-volume subtasks like routing, classification, and small multi-file code changes |
| 1M token context, 128K output | A large jump from Haiku 4.5's 200K context and 64K output limit, with a 300K output beta in Message Batches |
| Adaptive thinking | Thinking is adaptive with a default effort of medium, tuned for short, repetitive work |
| Computer and browser use | Well suited to repetitive browser and desktop tasks, with SDK support for computer use and browser use in beta |
Tokenizer
Haiku 5.5 uses an updated tokenizer, similar to Sonnet 5.5's and Opus 5.5's, which produces more tokens per task than the previous Haiku tokenizer.
Benchmark Performance
Independent evaluations · Artificial Analysis
Accuracy & Capability Details
Compare Models Side-by-Side
Evaluate specifications, pricing, and independent benchmark indices
| Model Details | |||
|---|---|---|---|
| General Info | |||
| Provider | Anthropic | Anthropic | Anthropic |
| Release Date | October 7, 2026 | September 22, 2026 | September 28, 2026 |
| Knowledge Cutoff | Jun 2026 | - | Jun 2026 |
| Context & Limits | |||
| Context Window | 1M | 1M | 1M |
| Pricing (per 1M tokens) | |||
| Input Pricing | $0.10 Best Input Pricing | $4 | $2 |
| Output Pricing | $0.50 Best Output Pricing | $20 | $10 |
| Modalities | |||
| Inputs | textimagefile | textimagefile | textimagefile |
| Outputs | text | text | text |
| Benchmarks (0-100) | |||
| Intelligence Index | 43.4 | 57.6 Best Intelligence Index | 56.0 |
| Coding Index | - | - | - |
| Agentic Index | - | - | - |
Humanity's Last Exam
Extremely difficult logical reasoning and knowledge.
Long Context Reasoning
Logical reasoning over long context windows.
SciCode Benchmark
Scientific coding and mathematical modeling.
Independent evaluation data provided by Artificial Analysis. To view the latest benchmarks and full details, visit their official site.
Explore more from Anthropic
Other models by Anthropic
Claude Opus 5.5
Claude Sonnet 5.5
Claude Fable 5.1
Claude Opus 5
Top AI Models
Leading alternatives by intelligence score
