Claude Sonnet Pricing:
Cost Per Token Calculator 2025
Claude Sonnet 4.5 (and 3.5) costs $3/M input · $15/M output. Calculate your exact session costs — paste your Claude Code log for a precise per-model breakdown.
Calculate My Costs →Claude Sonnet 4.5 — Official Pricing (2025)
| Token Type | Price per 1M Tokens | Notes |
|---|---|---|
| Input (standard) | $3.00 | Fresh context on every call |
| Output | $15.00 | Generated tokens |
| Cache write | $3.75 | First call that caches a block |
| Cache read Best value | $0.30 | Subsequent calls hit cache — 10× cheaper |
Claude Sonnet vs Other Models — Quick Comparison
Claude Sonnet sits in the mid-tier: smarter than Haiku, cheaper than Opus. Here's how it stacks up against common alternatives:
What Does Claude Sonnet Cost in Practice?
At $3/M input and $15/M output, Claude Sonnet handles complex reasoning while staying cost-efficient for most production workloads. Typical costs:
- 1 hour of Claude Code (typical session) — ~80K input + 20K output → $0.24 + $0.30 = $0.54
- Full workday of heavy Claude Code use (8h) — ~600K input + 150K output → $1.80 + $2.25 = ~$4.05
- With prompt caching (long system prompt) — same session, 60% cache hit → bill drops to ~$1.80–$2.20
- 1,000 document summaries (2K in, 300 out) → $6.00 + $4.50 = $10.50
Frequently Asked Questions
How much does Claude Sonnet cost per token in 2025?
Claude Sonnet 4.5 costs $3.00 per million input tokens and $15.00 per million output tokens. Prompt cache writes are $3.75/M and cache reads are $0.30/M — making repeated long-prompt calls up to 90% cheaper after the first call.
Is Claude Sonnet 4.5 cheaper than GPT-4o?
Claude Sonnet 4.5 costs slightly more per input token than GPT-4o ($3 vs $2.50), but Sonnet's cache-read price of $0.30/M beats GPT-4o's $1.25/M cached input price. For Claude Code sessions — which reuse large system prompts — Sonnet can end up cheaper than GPT-4o after caching kicks in.
What is the difference between Claude Sonnet 3.5 and Sonnet 4.5 pricing?
Both models share identical pricing: $3/M input, $15/M output. Claude Sonnet 4.5 offers stronger agentic capabilities and better multi-step reasoning, making it Anthropic's recommended default for Claude Code. There is no cost reason to stay on 3.5 unless you have compatibility requirements.
How do I calculate my exact Claude Sonnet costs from session logs?
Claude Code writes session logs to ~/.claude/projects/<project>/*.jsonl. Each entry records the model, token counts (input, output, cache write, cache read), and timestamp. Paste the file into the calculator above — it applies correct per-model pricing and breaks down the bill by model, tool call, and hour. Sonnet typically drives 80–95% of total Claude Code session cost.