Claude Sonnet 4.5 vs Prism
Claude Sonnet 4.5 is cheaper on output tokens. Choose Claude Sonnet 4.5 or Prism based on the trade-off between cost, context, and the benchmarks that matter for your use case.
| Spec | Claude Sonnet 4.5 | Prism |
|---|---|---|
| Provider | Anthropic | Charm Hyper |
| Input / 1M tokens | $3.00 | $3.27 |
| Output / 1M tokens | $15.00 | $16.33 |
| Context window | 1M | 1M |
| Parameters | — | — |
| Open weights | No | No |
| Released | Sep 2025 | Jan 1970 |
| LiveBench score | 67.9 | — |
| SWE-bench Multilingual score | 67.0 | — |
| GPQA Diamond score | 82.3 | — |
| FrontierMath score | 15.2 | — |
| SWE-bench Verified score | 74.8 | — |
| Epoch Capabilities Index score | 147 | — |
| LMArena (Chatbot Arena) Elo score | 1438 | — |
| LMArena WebDev Elo score | 1393 | — |
FAQ
Is Claude Sonnet 4.5 or Prism cheaper?
Claude Sonnet 4.5 is cheaper on output tokens ($15.00 vs $16.33 per 1M).
Which has the larger context window, Claude Sonnet 4.5 or Prism?
Both models offer a similar context window.
Pricing is indicative — confirm with the provider before production use. Highlighted values indicate the better figure for that row.