LLM APIFree tier
Claude 3.5 Sonnet · Anthropic
Quick answer
Claude 3.5 Sonnet is Anthropic's balanced workhorse model, strong at coding, analysis, and long documents. It is made by Anthropic, priced as “pay per token (from $3 / 1m input tokens)”, with observed latency around 290 ms and does not offer EU hosting on standard plans. Rated 4.8/5 by our editorial review.
Overview
Claude 3.5 Sonnet pairs frontier-adjacent quality with a 200K context window and notably strong writing and coding. Teams choose it when output quality, nuance, and safety matter more than the absolute lowest latency.
Benchmarks
| MMLU (5-shot) | 88.3% |
|---|---|
| HumanEval (coding) | 92.0% |
| Median latency (observed) | ~290 ms TTFT |
| Context window | 200K tokens |
Pros
- Excellent coding and long-document analysis
- 200K token context window
- Consistent, well-structured outputs
Cons
- No EU data residency on the standard API
- Occasional refusals on benign prompts