Claude Haiku 3.5 vs Claude Haiku 4.5: Pricing Comparison
Compare pricing, capabilities, and costs for your LLM workloads.
Anthropic
Claude Haiku 3.5
Pricing (per 1M tokens)
Input$0.8000
Output$4.00
Cached Input$0.0800
Batch Input$0.4000
Batch Output$2.00
Context & Output
Context Window200K tokens
Max Output8.2K tokens
Capabilities
Categorybudget
Multimodaltext + image
Fine-tuningNo
StreamingYes
Anthropic
Claude Haiku 4.5
Pricing (per 1M tokens)
Input$1.00
Output$5.00
Cached Input$0.1000
Batch Input$0.5000
Batch Output$2.50
Context & Output
Context Window200K tokens
Max Output64K tokens
Capabilities
Categorybudget
Multimodaltext + image
Fine-tuningNo
StreamingYes
Quick Verdict
Cheaper Input Price
Claude Haiku 3.5
20.0% cheaper
Cheaper Output Price
Claude Haiku 3.5
20.0% cheaper
Larger Context Window
Claude Haiku 4.5
+0 tokens
Cost Comparison
Sample workload: 1,000,000 input tokens + 1,000,000 output tokens
Claude Haiku 3.5
$4.80
$0.8000/1M input + $4.00/1M output
Claude Haiku 4.5
$6.00
$1.00/1M input + $5.00/1M output
Claude Haiku 3.5 is 20.0% cheaper for this workload.
Frequently Asked Questions
Which is cheaper, Claude Haiku 3.5 or Claude Haiku 4.5?
For input tokens, Claude Haiku 3.5 is cheaper at $0.8000 per 1M tokens. For output tokens, Claude Haiku 3.5 is cheaper at $4.00 per 1M tokens. The overall cost depends on your workload's input/output ratio.
What is the context window size of Claude Haiku 3.5 vs Claude Haiku 4.5?
Claude Haiku 3.5 has a context window of 200K tokens, while Claude Haiku 4.5 has 200K tokens. Claude Haiku 3.5 supports a larger context window of 200K tokens, which is beneficial for processing longer documents.
How do Claude Haiku 3.5 and Claude Haiku 4.5 compare for batch processing?
Both models support batch processing with discounted rates. Claude Haiku 3.5 offers a better batch rate at $0.4000 per 1M input tokens. Batch processing is ideal for non-time-sensitive workloads where you can wait for processing.
Need more tools?
Explore our complete suite of LLM calculators and comparison tools.