Claude 3.5 Haiku by Anthropic: fastest 3.5-family model with Claude 3 Opus-level performance, 40.6% SWE-bench, improved tool use, and 200K context window.
Capabilities, design details, and architectural traits
Claude 3.5 Haiku is Anthropic's fastest model in the 3.5 family, built around a specific trade-off: it delivers performance matching Claude 3 Opus on many evaluations while running at speeds comparable to the prior-generation Claude 3 Haiku. This generational jump - small model speed, large model capability - defines its identity.
| Trait | Detail |
|---|---|
| Coding performance | Scores 40.6% on SWE-bench Verified, surpassing the original Claude 3.5 Sonnet and GPT-4o on this task |
| Documented primary use cases | User-facing products, specialized sub-agent tasks, personalized experience generation from large data volumes |
| Tool use | Improved accuracy over Claude 3 Haiku; documented gains on agentic tool use evaluations |
| Instruction following | Explicitly improved over prior Haiku generation per official model card addendum |
Anthropic explicitly positions Claude 3.5 Haiku for sub-agent roles inside larger pipelines - fast, accurate tool calls with low latency. It is also documented as suited for generating personalized experiences from large structured datasets such as purchase history, pricing, or inventory records - a specific, non-generic use pattern that separates it from its Haiku predecessor.
Independent evaluations · Artificial Analysis
Evaluate specifications, pricing, and independent benchmark indices
| Model Details | |||
|---|---|---|---|
| General Info | |||
| Provider | Anthropic | Anthropic | Anthropic |
| Release Date | October 22, 2024 | July 24, 2026 | June 9, 2026 |
| Knowledge Cutoff | Jul 2024 | May 2026 | - |
| Context & Limits | |||
| Context Window | 200K | 1M Best Context Window | 1M Best Context Window |
| Pricing (per 1M tokens) | |||
| Input Pricing | $0.80 Best Input Pricing | $5 | $10 |
| Output Pricing | $4 Best Output Pricing | $25 | $50 |
| Modalities | |||
| Inputs | textimage | textimage | textimagefile |
| Outputs | text | text | text |
| Benchmarks (0-100) | |||
| Intelligence Index | 12.2 | 63.1 Best Intelligence Index | 62.1 |
| Coding Index | 15.9 | 78.0 Best Coding Index | 76.5 |
| Agentic Index | 1.9 | 59.2 Best Agentic Index | 56.6 |
Graduate-level reasoning and expert Q&A evaluation.
Extremely difficult logical reasoning and knowledge.
Logical reasoning over long context windows.
Independent evaluation data provided by Artificial Analysis. To view the latest benchmarks and full details, visit their official site.
