What is the cheapest AI API for building a chatbot?

For chatbots, DeepSeek V4 Flash ($0.14/$0.28 per million tokens) is the cheapest option that maintains good conversation quality. A typical 10-turn chatbot conversation costs ~$0.0003 with DeepSeek V4 Flash vs $0.0038 with GPT-5 — a 92% savings. For basic FAQ bots where quality is less critical, Llama 3.1 8B ($0.10/$0.10) is even cheaper.

Which AI API is cheapest for code generation?

DeepSeek V4 Pro ($0.435/$0.87) is the best value for code generation — it matches GPT-5.3 Codex quality at 93% lower output cost. A typical code generation request (2K input, 4K output) costs $0.0039 with DeepSeek V4 Pro vs $0.058 with GPT-5.3 Codex. For simple completions, Gemini 3 Flash ($0.50/$3.00) is also competitive.

What is the cheapest AI API for RAG (Retrieval-Augmented Generation)?

RAG workloads are output-heavy, making DeepSeek V4 Pro ($0.435/$0.87) the cheapest quality option — a typical RAG query (10K context + 1K output) costs $0.0052. For high-volume RAG where you can tolerate slightly lower quality, Gemini 2.5 Flash-Lite ($0.10/$0.40) costs just $0.0008 per query — 84% cheaper than DeepSeek. The key is matching model quality to your RAG accuracy requirements.

Which AI API is cheapest for text summarization?

Summarization is input-heavy (long documents in, short summary out). Gemini 2.5 Flash-Lite ($0.10/$0.40) is the cheapest at $0.0005 per 5,000-word document. DeepSeek V4 Flash ($0.14/$0.28) is close at $0.0007. For high-quality summaries of complex documents, Claude Haiku 4.5 ($1.00/$5.00) costs $0.006 but delivers significantly better coherence.

What is the cheapest AI API for embeddings?

For dedicated embeddings, OpenAI's text-embedding-3-small ($0.02/1M tokens) and Cohere's embed-english-v3.0 ($0.10/1M tokens) are the standard choices. However, several providers now offer free or near-free embedding endpoints. For vector search at scale, Mistral Small 4 ($0.15/$0.60) can handle embedding + generation in a single call, eliminating a separate embedding API cost.

How do I choose the cheapest AI API for my specific use case?

The cheapest API depends on your input/output token ratio, quality requirements, and volume. Use the APIpulse Model Recommendation Engine — answer 4 questions about your use case, and it recommends the top 3 models with projected monthly costs. For exact calculations, the Cost Calculator lets you model your token usage across all 88 models.