Model Comparison
Claude 3.5 Haiku
vs. Claude 3.7 Sonnet (Non-reasoning)
Comparing 2 AI models · 12 benchmarks · Anthropic
Recommended Pick
An Claude 3.7 Sonnet (Non-reasoning) 14 metric wins Strongest on: Reasoning, Intelligence, Coding
Best Value
Claude 3.5 Haiku
100.0 value score
26.4 reasoning / $1.60/1M
Lowest Price
Claude 3.5 Haiku
$0.80/1M input price
Best Reasoning
Claude 3.7 Sonnet (Non-reasoning)
34.7 reasoning score
Blends available reasoning benchmarks
Best for Coding
Claude 3.7 Sonnet (Non-reasoning)
26.7 coding index
Composite Indices
Higher is better; speed and price are normalized
Standard Benchmarks
Only benchmarks with data are shown
Differences That Matter
Best value
Claude 3.5 Haiku has the strongest quality-to-price mix at 100.0 out of 100 value points.
Price gap
Claude 3.5 Haiku is 3.8x cheaper on input tokens than Claude 3.7 Sonnet (Non-reasoning).
Reasoning gap
Claude 3.7 Sonnet (Non-reasoning) leads Claude 3.5 Haiku by 8.3 points on reasoning.
Coding gap
Claude 3.7 Sonnet (Non-reasoning) leads Claude 3.5 Haiku by 16.0 points on coding.
Top-pick rationale
Claude 3.7 Sonnet (Non-reasoning) wins 14 measurable categories, including Reasoning, Intelligence, Coding, GPQA.
Response Face-Off
Run one prompt through the selected models and compare response quality with live speed and cost context.
Claude 3.5 Haiku
Anthropic
TTFT
—
Time
—
tok/s
—
Tokens
—
Cost
—
Claude 3.7 Sonnet (Non-reasoning)
Anthropic
TTFT
—
Time
—
tok/s
—
Tokens
—
Cost
—
Which answer was more useful?
Full Comparison
| Metric | An Claude 3.5 Haiku | Top Pick An Claude 3.7 Sonnet (Non-reasoning) |
|---|---|---|
| Pricing per 1M tokens | ||
| Input Cost | $0.80/1M | $3.00/1M |
| Output Cost | $4.00/1M | $15.00/1M |
| Blended (3:1) | $1.60/1M | $6.00/1M |
| Specifications | ||
| Organization | Anthropic | Anthropic |
| Release Date | Oct 22, 2024 | Feb 24, 2025 |
| Performance & Speed | ||
| Throughput | — | — |
| TTFT | — | — |
| Latency | — | — |
| Composite Indices | ||
| Value Score | 100.0 | 35.1 |
| Reasoning Score | 26.4 | 34.7 |
| Intelligence | 12.4 | 23.5 |
| Coding | 10.7 | 26.7 |
| Math | — | 21.0 |
| Standard Benchmarks | ||
| GPQA | 40.8% | 65.6% |
| MMLU Pro | 63.4% | 80.3% |
| HLE | 3.5% | 4.8% |
| LiveCodeBench | 31.4% | 39.4% |
| MATH 500 | 72.1% | 85.0% |
| AIME 2025 | — | 21.0% |
| AIME (Original) | 3.3% | 22.3% |
| SciCode | 27.4% | 37.6% |
| LCR | 23.3% | 48.3% |
| IFBench | 42.8% | 44.0% |
| TAU-bench v2 | 24.6% | 50.0% |
| TerminalBench Hard | 2.3% | 21.2% |
Key Takeaways
Claude 3.5 Haiku offers the best value at $0.80/1M, making it ideal for high-volume applications and cost-conscious projects.
Claude 3.7 Sonnet (Non-reasoning) has the strongest reasoning profile with a 34.7 reasoning score, combining the available reasoning-heavy benchmarks.
Claude 3.7 Sonnet (Non-reasoning) reaches a 26.7 coding index, making it the top choice for software development and code generation tasks.
All models support context windows of ∞+ tokens, suitable for processing lengthy documents and maintaining extended conversations.
When to Choose Each Model
Claude 3.5 Haiku
- Cost-sensitive applications
- High-volume processing
Claude 3.7 Sonnet (Non-reasoning)
- Complex reasoning tasks
- Research & analysis
- Code generation
- Software development
