Cheapest LLM APIs, ranked
Ranked by cost for a typical chat call (2K input / 500 output tokens). Cheapest is not always best value: check quality on your own prompts.
| # | Model | Chat 2K/500 | RAG 20K/1K | Long out 500/4K |
|---|---|---|---|---|
| 1 | Gemini 2.5 Flash-Lite | $0.0004 | $0.0024 | $0.00165 |
| 2 | GPT-6 Luna | $0.00045 | $0.0025 | $0.00205 |
| 3 | Claude Haiku 5.5 | $0.00045 | $0.0025 | $0.00205 |
| 4 | GPT-4o mini | $0.0006 | $0.0036 | $0.00248 |
| 5 | GPT-5.6 Luna | $0.001 | $0.0052 | $0.0049 |
| 6 | GPT-5.4 nano | $0.00103 | $0.00525 | $0.0051 |
| 7 | Gemini 3.1 Flash-Lite | $0.00125 | $0.0065 | $0.00613 |
| 8 | GPT-5 mini | $0.0015 | $0.007 | $0.00813 |
| 9 | Gemini 3.5 Flash-Lite | $0.00185 | $0.0085 | $0.0101 |
| 10 | Gemini 2.5 Flash | $0.00185 | $0.0085 | $0.0101 |
| 11 | Grok Build 0.1 | $0.003 | $0.0220 | $0.0085 |
| 12 | Gemini 3.8 Flash | $0.00337 | $0.0187 | $0.0154 |
| 13 | Gemini 3.6 Flash | $0.00337 | $0.0187 | $0.0154 |
| 14 | GPT-5.4 mini | $0.00375 | $0.0195 | $0.0184 |
| 15 | Grok 4.3 | $0.00375 | $0.0275 | $0.0106 |
| 16 | o4-mini | $0.0044 | $0.0264 | $0.0181 |
| 17 | Claude Haiku 4.5 | $0.0045 | $0.0250 | $0.0205 |
| 18 | Grok 4.7 | $0.007 | $0.0460 | $0.0250 |
| 19 | Gemini 2.5 Pro | $0.0075 | $0.0350 | $0.0406 |
| 20 | GPT-4.1 | $0.008 | $0.0480 | $0.0330 |
| 21 | GPT-6.1 Sol | $0.009 | $0.0500 | $0.0410 |
| 22 | GPT-6 Sol | $0.009 | $0.0500 | $0.0410 |
| 23 | Claude Sonnet 5.5 | $0.009 | $0.0500 | $0.0410 |
| 24 | Claude Sonnet 5 | $0.009 | $0.0500 | $0.0410 |
| 25 | Gemini 4 Argon | $0.009 | $0.0500 | $0.0410 |
| 26 | GPT-5.6 Terra | $0.0100 | $0.0520 | $0.0490 |
| 27 | Gemini 3.1 Pro Preview | $0.0100 | $0.0520 | $0.0490 |
| 28 | GPT-5.4 | $0.0125 | $0.0650 | $0.0612 |
| 29 | Claude Sonnet 4.6 | $0.0135 | $0.0750 | $0.0615 |
| 30 | GPT-5.6 Sol | $0.0180 | $0.1000 | $0.0820 |
| 31 | Claude Opus 5.5 | $0.0180 | $0.1000 | $0.0820 |
| 32 | Claude Opus 5 | $0.0225 | $0.1250 | $0.1025 |
| 33 | GPT-5.5 | $0.0250 | $0.1300 | $0.1225 |
| 34 | GPT-6 Astra | $0.0450 | $0.2500 | $0.2050 |
| 35 | Claude Fable 5.1 | $0.0450 | $0.2500 | $0.2050 |
Per single call, USD, standard tier, no caching. Several low-cost rates are promotional or capped by prompt length; see price changes.
Next step → Compare the cheapest models for your own token counts
Prices last checked 2026-10-26 against official sources · how we verify
Price-change alerts
One short email when a provider changes a price or a promo ends. No spam, unsubscribe any time.