Head to head
Anthropic: Claude Sonnet 5 vs Google: Gemini 3.6 Flash
Live prices, context windows, capabilities, and measured quality — updated continuously.
Side by side
| Metric | Anthropic: Claude Sonnet 5 Anthropic | Google: Gemini 3.6 Flash Google |
|---|---|---|
| Input $/M tokens | $2.00 | $0.750 |
| Output $/M tokens | $10.00 | $3.75 |
| Context window | 1.0M | 1.0M |
| Max output | 128K | 66K |
| WM Score | 87 | 83 |
| Benchmark quality | 85.5 | 80.3 |
| Hosts serving it | 5 | 2 |
| Throughput (tok/s) | 73 | 112 |
| Latency (ms) | 1182 | 1741 |
| Vision | ||
| Tools | ||
| Reasoning | ||
| JSON mode |
Which should you pick?
Google: Gemini 3.6 Flash costs less per million input tokens — about 2.7× cheaper than Anthropic: Claude Sonnet 5. Anthropic: Claude Sonnet 5 scores higher on measured quality. Google: Gemini 3.6 Flash takes a larger context window, which matters for long documents or big codebases. Prices here are live list prices; the same model can cost less on a different host, so check the model page for the cheapest offering.