AI Model Comparison Calculator
Enter your expected monthly token volume and see the estimated cost across current Claude, GPT, and Gemini models side by side — plus each model's context window and max output, so you can compare capability, not just price.
Compare all models
Compare all models
| Model | Provider | Context window | Max output | $/1M in | $/1M out | Est. monthly cost |
|---|---|---|---|---|---|---|
| GPT-5.6 Luna | OpenAI | 1.05M | 128K | $0.20 | $1.20 | $4.40 |
| Gemini 3.5 Flash-Lite | 1.05M | 66K | $0.30 | $2.50 | $8.00 | |
| Gemini 3.6 Flash | 1.05M | 66K | $0.75 | $3.75 | $15.00 | |
| Claude Haiku 4.5 | Anthropic | 200K | 64K | $1.00 | $5.00 | $20.00 |
| Claude Sonnet 5 | Anthropic | 1M | 128K | $2.00 | $10.00 | $40.00 |
| GPT-5.6 Terra | OpenAI | 1.05M | 128K | $2.00 | $12.00 | $44.00 |
| Gemini 3.1 Pro Preview | 1.05M | 66K | $2.00 | $12.00 | $44.00 | |
| GPT-5.6 Sol | OpenAI | 1.05M | 128K | $4.00 | $20.00 | $80.00 |
| Claude Opus 5 | Anthropic | 1M | 128K | $5.00 | $25.00 | $100.00 |
Gemini 3.1 Pro Preview's real pricing roughly doubles for the portion of any SINGLE request beyond 200K tokens of context — a per-request tier, not a monthly one. This calculator has no way to know how your monthly volume splits across individual requests, so it always estimates using the base (≤200K) rate. If you regularly send requests with more than 200K tokens of context, your real Gemini 3.1 Pro Preview cost will be higher than shown.
Prices and limits verified against each provider's own current pricing page on 2026-08-25 — subscription/API prices change; check the provider's current page before committing to usage. Gemini 3.6 Flash's current price is explicitly marked by Google as valid "through Dec 31, 2026" (promotional) — verify it again after that date.
How this is calculated
Monthly cost for each model is your entered input-token volume × that model's price per million input tokens, plus your output-token volume × its price per million output tokens — standard, non-batch, non-cached API pricing for every model, so the comparison is fair across providers. This does NOT model per-request context-length pricing tiers (see the note next to Gemini 3.1 Pro Preview), batch API or prompt-caching discounts, or multimodal (image/audio) token costs — it estimates plain text input/output at the standard rate only.
Common questions
Why does Gemini 3.1 Pro Preview have a warning about 200K tokens?
Google prices that model per SINGLE request: requests with more than 200K tokens of context cost roughly double per token. This calculator only knows your total monthly volume, not how it's split across individual requests, so it always estimates using the cheaper (≤200K) rate — your real cost will be higher if you regularly send large-context requests.
Does this account for batch API or prompt caching discounts?
No — every model is estimated at its standard, non-batch, non-cached API rate. Batch processing and prompt caching can meaningfully lower real costs for some workloads; this calculator intentionally doesn't model either, to keep the comparison consistent across all three providers.
Where do the prices come from, and how often are they updated?
Prices are checked manually against each provider's own public pricing page and stored with the date they were verified (shown below the table) — this calculator doesn't poll provider APIs live, so always check the provider's current page before committing to a usage plan.
Why are only Claude, GPT, and Gemini models included?
This first version covers the three most commonly compared providers. Other providers may be added later, but aren't part of this catalog today.
Part of Dev & AI Tool Calculators