LLM prices / Cheapest / RAG
AI API prices · 29 models fit · checked Oct 23, 2026
Cheapest LLM for RAG
Short answerClaude Haiku 5.5 and GPT-6 Luna tie for the cheapest current model for a RAG answer at $0.85 per 1,000, standard price. At 30,000 a month that is $25.50. Next: GPT-5.6 Luna at $1.80 and Gemini 3.1 Flash-Lite at $2.25 per 1,000. 29 of 29 current models fit this job. Ranked by cost only; we don't score quality.
Every model that fits, cheapest first
One RAG answer = 6,000 tokens of retrieved passages plus the question, and a 500-token answer. Each model is billed the cheapest way its price list allows.
| # | Model | Per 1,000 | 30,000 a month | 300,000 a month | Billed | Context |
|---|---|---|---|---|---|---|
| 1 | Claude Haiku 5.5Anthropic | $0.85 | $25.50 | $255 | standard | 1M |
| 2 | GPT-6 LunaOpenAI | $0.85 | $25.50 | $255 | standard | 1.05M |
| 3 | GPT-5.6 LunaOpenAI | $1.80 | $54.00 | $540 | standard | 1.05M |
| 4 | Gemini 3.1 Flash-LiteGoogle | $2.25 | $67.50 | $675 | standard | 1.05M |
| 5 | DeepSeek FlashDeepSeek | $2.40 | $72.00 | $720 | standard | 1M |
| 6 | Gemini 3.5 Flash-LiteGoogle | $3.05 | $91.50 | $915 | standard | 1.05M |
| 7 | Gemini 3 Flash PreviewpreviewGoogle | $4.50 | $135 | $1,350 | standard | 1.05M |
| 8 | Gemini 3.8 FlashGoogle | $6.38 | $191 | $1,912 | standard | 1.05M |
| 9 | Gemini 3.6 FlashGoogle | $6.38 | $191 | $1,912 | standard | 1.05M |
| 10 | GPT-5.4 miniOpenAI | $6.75 | $202 | $2,025 | standard | 400K |
| 11 | DeepSeek V4 ProDeepSeek | $9.90 | $297 | $2,970 | standard | 1M |
| 12 | Claude Sonnet 5.5Anthropic | $17.00 | $510 | $5,100 | standard | 1M |
| 13 | GPT-6.1 SolOpenAI | $17.00 | $510 | $5,100 | standard | 1.05M |
| 14 | GPT-6 SolOpenAI | $17.00 | $510 | $5,100 | standard | 1.05M |
| 15 | Gemini 4 ArgonGoogle | $17.00 | $510 | $5,100 | standard | 1.05M |
| 16 | GPT-5.2OpenAI | $17.50 | $525 | $5,250 | standard | 400K |
| 17 | GPT-5.6 TerraOpenAI | $18.00 | $540 | $5,400 | standard | 1.05M |
| 18 | Gemini 3.1 Pro PreviewpreviewGoogle | $18.00 | $540 | $5,400 | standard | 1.05M |
| 19 | GPT-5.4OpenAI | $22.50 | $675 | $6,750 | standard | 1.05M |
| 20 | Claude Opus 5.5Anthropic | $34.00 | $1,020 | $10,200 | standard | 1M |
| 21 | GPT-5.6 SolOpenAI | $34.00 | $1,020 | $10,200 | standard | 1.05M |
| 22 | GPT-5.5OpenAI | $45.00 | $1,350 | $13,500 | standard | 1.05M |
| 23 | chat-latestOpenAI | $45.00 | $1,350 | $13,500 | standard | 400K |
| 24 | Claude Fable 5.1Anthropic | $85.00 | $2,550 | $25,500 | standard | 1M |
| 25 | GPT-6 AstraOpenAI | $85.00 | $2,550 | $25,500 | standard | 1.05M |
| 26 | GPT-5.6 CyberOpenAI | $112 | $3,375 | $33,750 | standard | 400K |
| 27 | GPT-5.2 ProOpenAI | $210 | $6,300 | $63,000 | standard | 400K |
| 28 | GPT-5.5 ProOpenAI | $270 | $8,100 | $81,000 | standard | 1.05M |
| 29 | GPT-5.4 ProOpenAI | $270 | $8,100 | $81,000 | standard | 1.05M |
What moves the answer
- Input price matters most here. The job reads 12x more than it writes, so the input (and cache-read) price drives the bill.
- Watch long prompts. 13 of these models charge more once a prompt passes a size threshold (for example Claude Haiku 5.5 above 100,000 tokens); this job stays under it, a bigger one may not.
Method
Prices: each provider's own pricing page, checked Oct 23, 2026. Fits: current on the provider's model page (not deprecated, legacy or shut down), and a context window that holds the request; checked Oct 24, 2026. Cache writes are not counted. This page ranks price only; quality on your task is a separate test. People search for this: Google suggests "cheapest llm for rag".