← All Comparisons
Claude Haiku 4.5 vs Claude Sonnet 4.6
A detailed comparison of Claude Haiku 4.5 (Anthropic) and Claude Sonnet 4.6 (Anthropic) across pricing, performance, and features.
Pricing Comparison
| Metric | Claude Haiku 4.5 | Claude Sonnet 4.6 | Difference |
|---|---|---|---|
| Input / 1M tokens | $0.80 | $3.00 | +275% |
| Output / 1M tokens | $4.00 | $15.00 | +275% |
| Context window | 200K | 200K | — |
| Max output | 8.192K | 16K | — |
Benchmark Comparison
| Benchmark | Claude Haiku 4.5 | Claude Sonnet 4.6 |
|---|---|---|
| MMLU-Pro | 69.4% | 86% |
| HumanEval | 88.1% | 94% |
| GPQA | — | 70% |
Capabilities
| Capability | Claude Haiku 4.5 | Claude Sonnet 4.6 |
|---|---|---|
| code | ✓ | ✓ |
| reasoning | ✗ | ✓ |
| text | ✓ | ✓ |
| tool-use | ✓ | ✓ |
| vision | ✓ | ✓ |
Claude Haiku 4.5 Strengths
- ✓Very fast responses
- ✓Cheapest Anthropic option
- ✓Good for classification and extraction
Claude Haiku 4.5 Weaknesses
- ✗Weakest reasoning in the Claude family
- ✗Can struggle with nuanced instructions
Claude Sonnet 4.6 Strengths
- ✓Opus 4.5 quality at 1/5th the cost
- ✓Best value for production workloads
- ✓1M context in beta
Claude Sonnet 4.6 Weaknesses
- ✗Long context pricing doubles above 200K
- ✗Slightly below Opus 4.6 on hardest tasks
Quick Verdict
Best value: Claude Haiku 4.5 is the more affordable option at $0.8/$4 per 1M tokens.
Higher benchmarks: Claude Sonnet 4.6 scores higher on average across available benchmarks (83.3% avg).
Choose Claude Haiku 4.5 if cost matters most. Choose Claude Sonnet 4.6 if you need the best possible quality for complex tasks.