Which AI API is cheapest?

GPT-4o mini and Gemini 2.5 Flash are tied for cheapest among tracked models at $0.15 input / $0.60 output per million tokens, as of the last verification date shown on this site.

How often is LLM Price Watch's pricing data updated?

Prices are verified directly against each provider's official documentation, not third-party aggregators, and re-checked regularly. The verification date is shown on every page.

Does batch processing or prompt caching change these prices?

Yes. Batch processing typically cuts costs by roughly 50%, and prompt caching can cut the cost of repeated context by roughly 90% on cache hits. Prices shown by default are standard, non-batch, non-cached list rates.

LLM Price Watch — Live AI API pricing across every major provider

MONTHLY COST CALCULATOR

Enter your real usage

Rough numbers are fine — most teams don't know their exact split until they've shipped. Input and output tokens are priced separately because output costs 3–6x more across every provider.

Input tokens per month (millions)

Output tokens per month (millions)

Model	Provider	Relative cost	Monthly cost

Calculated from standard (non-batch, non-cached) list pricing. Real bills are usually lower once caching and batch discounts are applied.

Embed this calculator on your own site →

<iframe src="https://llmpricewatch.com/embed/calculator" width="100%" height="420" frameborder="0"></iframe>

Free to embed anywhere. Keeps a small link back to llmpricewatch.com for the full comparison.

Read the full breakdown: Claude vs GPT vs Gemini pricing by tier →

Read next: AI API free tiers compared →

Comparing raw API pricing here, not which tool to actually build with — StackIndex AI covers the SaaS and AI tools themselves for solo founders.

FULL PRICE TABLE

Every tracked model, side by side

Sorted by output price, low to high. Verified against each provider's official pricing page — sources listed in the methodology below.

Model	Provider	Tier	Input / 1M	Output / 1M	Context

PROVIDER GUIDES

Deep dive by provider

Full pricing breakdowns for each provider we track — tier-by-tier, with the trade-offs that matter beyond the sticker price.

Anthropic Claude

4 tiers, and one hard price-change date to know about.

OpenAI GPT

The widest price spread of any provider we track.

Google Gemini

Why context window size matters more than the per-token rate.

HEAD-TO-HEAD

Same tier, different provider

Direct comparisons between models in the same price class, so you're comparing like for like.

Claude Haiku 4.5 vs GPT-4o mini Claude Haiku 4.5 vs Gemini 2.5 Flash GPT-4o mini vs Gemini 2.5 Flash Claude Sonnet 5 vs GPT-4o Claude Sonnet 5 vs Gemini 3 Flash Claude Opus 4.8 vs GPT-5.5 Claude Opus 4.8 vs Gemini 3.1 Pro Claude Opus 4.8 vs Sonnet 5 Gemini 3 Flash vs GPT-4o mini DeepSeek V3 vs Grok 4.1

METHODOLOGY

How we verify these numbers

Every price on this page is read directly from the provider's official API pricing documentation — not from a reseller, not from a third-party aggregator. Prices shown are standard, pay-as-you-go list rates per million tokens; they exclude prompt-caching discounts (typically 90% off on cache hits) and batch-processing discounts (typically 50% off), both of which can meaningfully lower a real production bill. Snapshot last verified 11 July 2026. API pricing changes often — if you're making a spending commitment over a few hundred dollars a month, re-check the provider's own page before you commit.

Sources: platform.claude.com/docs/about-claude/pricing, platform.openai.com/docs/pricing, ai.google.dev/gemini-api/docs/pricing, api-docs.deepseek.com, docs.x.ai.

Get an email when a tracked model's price changes

One email per price change. Nothing else.

Know what your AI bill will actually be, before it lands.