Claude 3 Haiku
Anthropic
Anthropic's fastest and most affordable model in its intelligence class. Processes 21K tokens (~30 pages) per second for prompts under 32K tokens, making it 3x faster than peers. Features 200K context window, vision capabilities, and surprisingly strong performance - during testing, surpassed the previous flagship Claude 3 Opus on many benchmarks. Exceptional value proposition: can process 400 Supreme Court cases or 2,500 images for just $1. State-of-the-art speed-to-intelligence ratio.
Strengths
- Exceptional speed - 3x faster than peers, processes 21K tokens (~30 pages) per second
- Best price-to-performance ratio - $0.25 per 1M input tokens
- 200K context window handles large documents and long conversations
- Vision capabilities with function calling support
- Surprising quality - surpassed Claude 3 Opus on many benchmarks during testing
- Can process 400 Supreme Court cases or 2,500 images for $1
- State-of-the-art performance for speed-optimized tier
Caveats
- Lower ceiling than Claude 3 Opus or Sonnet for most complex tasks
- August 2023 knowledge cutoff older than competitors
- 4K max output tokens lower than newer models
- Not ideal for tasks requiring maximum reasoning depth
- Vision capabilities less sophisticated than larger models
- Vendor
- Anthropic
- Context Window
- 200,000 tokens
- Max Output
- 4,096 tokens
- License
- Proprietary
- Release Date
- 2024-03-13
- Training Cutoff
- 2023-08-01
- Modalities
- textimage
| Benchmark | Score | Max |
|---|---|---|
| livebench_language | 29.6 | 100.0 |
| livebench_coding | 24.5 | 100.0 |
| livebench_instruction_following | 57.0 | 100.0 |
Capabilities
Vision
Audio
Video
Tool Use
Resources
Pricing
$0.25 per 1M input tokens, $1.25 per 1M output tokens