Claude 3 Haiku
Claude 3 Haiku by Anthropic: a compact Claude 3 variant optimized for near-instant responses, 200K-token context window, 4K max output, Aug 2023 cutoff.
Model Overview
Capabilities, design details, and architectural traits
Claude 3 Haiku - Anthropic compact, low-latency model
Claude 3 Haiku is the compact member of Anthropic's Claude 3 family designed for near-instant responses and high efficiency. It trades top-end generative depth for lower latency and smaller runtime footprint while keeping the large context window used across the Claude 3 family.
| Trait | Documented detail that distinguishes Claude 3 Haiku |
|---|---|
| Identity within family | The smallest and fastest Claude 3 variant in Anthropic's Claude 3 lineup (Haiku vs Sonnet vs Opus). |
| Optimization target | Explicitly optimized for near-instant response speed and runtime efficiency. |
| Context capacity | Uses the Claude 3 family context window of up to 200,000 tokens/characters as documented for Haiku. |
| Output limit | Documented maximum single response output capped at 4,000 characters/tokens. |
| Primary positioning | Positioned for use cases needing low latency and lower compute cost compared with larger Claude 3 models. |
| Official availability | Offered through Anthropic channels and available on cloud marketplaces (for example, Amazon Bedrock) as the Haiku edition. |
Notable model-specific points
- Compact-speed tradeoff: Haiku is explicitly presented as the compact, low-latency sibling within the Claude 3 family rather than the family’s highest-capability variant.
- Same large context approach: Despite its compact runtime goal, Haiku retains the Claude 3 family’s very large context capability, making it suitable for applications that need long context with fast responses.
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 | March 4, 2024 | September 22, 2026 | September 28, 2026 |
| Knowledge Cutoff | Aug 2023 | - | Jun 2026 |
| Context & Limits | |||
| Context Window | 200K | 1M Best Context Window | 1M Best Context Window |
| Pricing (per 1M tokens) | |||
| Input Pricing | $0.25 Best Input Pricing | $4 | $2 |
| Output Pricing | $1.25 Best Output Pricing | $20 | $10 |
| Modalities | |||
| Inputs | textimage | textimagefile | textimagefile |
| Outputs | text | text | text |
| Benchmarks (0-100) | |||
| Intelligence Index | 5.6 | 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
