Cheapest AI API 2026:
Lowest Cost LLMs Ranked

Cheapest AI APIs Ranked by Input Price (July 2026)

Prices are per million tokens (1M tokens ≈ 750,000 words). Sorted by input token cost, lowest first. Current models only — retired models (Claude 3.x, Gemini 1.5/2.0) have been removed.

Top 3 Cheapest AI APIs — Detailed Breakdown

🥇 #1: Gemini 2.5 Flash-Lite — $0.10/M input

Google's most affordable model in 2026. Flash-Lite is designed for simple, repetitive tasks at extreme scale. It handles classification, extraction, summarization, and basic Q&A well. Not suitable for complex multi-step reasoning or nuanced writing.

• Best for: Content moderation, data extraction, simple classification, FAQ bots
• Avoid for: Creative writing, complex coding, legal/medical analysis
• 1M token cost: $0.10 input + $0.40 output = $0.50 total for 500K in / 500K out split

🥈 #2: GPT-4o mini (legacy) — $0.15/M input

OpenAI's legacy budget model still punches above its price and remains callable. It delivers strong performance on structured tasks, follows instructions reliably, and integrates with OpenAI's tools ecosystem. If you want the current-generation equivalent, GPT-5.4 mini ($0.75/$4.50) adds newer capability at a higher floor.

• Best for: High-volume customer support, data processing, content generation at scale
• Avoid for: Tasks requiring frontier-level reasoning (use GPT-5.4 or Claude Sonnet 4.6 instead)
• Sweet spot: Applications where you'd use a flagship but cost is the constraint

🥉 #3: Gemini 2.5 Flash — $0.30/M input

A step up from Flash-Lite with a massive 1M token context window — making it a strong budget choice for long-document processing. Summarizing a 200-page PDF costs only a few cents in input tokens.

• Best for: Long document analysis, RAG pipelines, multi-document summarization
• Unique advantage: 1M context window at budget pricing

How Much Can You Save by Switching Models?

Assume a customer support chatbot processing 1 million requests/month with 500 input + 200 output tokens each:

Should You Use Open-Source Models (Llama, Mistral)?

Meta's Llama and Mistral's open-weight models can be significantly cheaper when self-hosted, but running your own infrastructure adds complexity and fixed costs — and hosted pricing varies by provider. AIModelCalc currently tracks the three first-party providers (OpenAI, Anthropic, Google) where published per-token pricing is stable; for open-weight models, check your hosting provider's current rate, since it's often competitive with the cheapest closed-source options above.

Self-hosting makes sense if you process more than 50M tokens/month and have engineering capacity. Below that threshold, hosted APIs are simpler and often cheaper when you factor in GPU costs and maintenance.

Tips for Minimizing AI API Costs

• Use prompt caching — Anthropic and OpenAI both offer 50–90% discounts on repeated input context
• Trim system prompts — Every token in your system prompt is charged on every request
• Match model to task — Use small models for simple tasks, save premium models for complex work
• Batch non-urgent requests — OpenAI's Batch API offers 50% off for async workloads
• Monitor output length — Output tokens cost 4–5x more than input; set max_tokens limits