LLM Cost Comparison 2026 — Claude, GPT, Gemini
Side-by-side per-million-token pricing for Claude Opus 4.7, Sonnet 4.6, Haiku 4.5, GPT-5 Turbo, and Gemini 2.5 Pro. Open dataset under CC BY 4.0.
Per-million-token pricing (USD, May 2026)
| Vendor | Model | Input | Output | Cached read | Context window |
|---|
| Anthropic | Claude Opus 4.7 | $15 | $75 | $1.5 | 200,000 |
| Anthropic | Claude Sonnet 4.6 | $3 | $15 | $0.3 | 200,000 |
| Anthropic | Claude Haiku 4.5 | $1 | $5 | $0.1 | 200,000 |
| OpenAI | GPT-5 Turbo | $5 | $20 | $2.5 | 256,000 |
| Google | Gemini 2.5 Pro | $2.5 | $10 | $0.625 | 2,000,000 |
Notes per model
- Claude Opus 4.7 — Top-tier reasoning, vision, tool use
- Claude Sonnet 4.6 — Recommended default for production
- Claude Haiku 4.5 — Volume classification and extraction
- GPT-5 Turbo — Public pricing snapshot 2026-05; verify on openai.com
- Gemini 2.5 Pro — Public pricing snapshot 2026-05; 2M context tier
Download the dataset
This page is published as an open dataset under CC BY 4.0. Attribution: Claude API Cookbook.
Caveats
Pricing reflects public vendor pages as of May 2026. Verify against the source before committing budget:
anthropic.com/pricing,
openai.com/pricing,
ai.google.dev/pricing.
Free tools
Related pages
🔥 First-10 launch: Claude Code Power Prompts PDF · £5 £0.50Get the PDF → 📘 30 prompts. Paste into CLAUDE.md. Never re-type a prompt. £0.50Buy the pack → 🛡 30 Claude Code prompts · £0.50 · 7-day refund, no questionsTry the PDF →
⏸ Before you go…
If the cookbook helped, the full Claude Code Power Prompts pack has 30 paste-into-CLAUDE.md templates. Pay what you can.
Pay what you want · from £2 →
8-page PDF · 30 prompts · 7-day refund