Cheapest LLM APIs
Ranked by a blended price that weights input and output 3 to 1, a common mix for chat and document tasks. At the prices we checked, the three lowest are GPT-5 nano ($0.138), GPT-4.1 nano ($0.175), Gemini 2.5 Flash-Lite ($0.175) per million blended tokens.
| Model | Provider | Input | Cached input | Output | Blended |
|---|---|---|---|---|---|
| GPT-5 nano | OpenAI | $0.050 | $0.0050 | $0.400 | $0.138 |
| GPT-4.1 nano | OpenAI | $0.100 | $0.025 | $0.400 | $0.175 |
| Gemini 2.5 Flash-Lite Input price shown is for text, image and video; audio input costs more. | Google (Gemini) | $0.100 | $0.010 | $0.400 | $0.175 |
| GPT-6 Luna | OpenAI | $0.100 | $0.010 | $0.500 | $0.200 |
| GPT-4o mini | OpenAI | $0.150 | $0.075 | $0.600 | $0.262 |
| GPT-5.6 Luna | OpenAI | $0.200 | $0.020 | $1.20 | $0.450 |
| GPT-5.4 nano | OpenAI | $0.200 | $0.020 | $1.25 | $0.463 |
| GPT-5 mini | OpenAI | $0.250 | $0.025 | $2.00 | $0.688 |
| GPT-4.1 mini | OpenAI | $0.400 | $0.100 | $1.60 | $0.700 |
| Gemini 2.5 Flash Input price shown is for text, image and video; audio input costs more. | Google (Gemini) | $0.300 | $0.030 | $2.50 | $0.850 |
| Gemini 3.8 Flash | Google (Gemini) | $0.750 → $1.50 from 2027-01-01 | $0.075 → $0.150 from 2027-01-01 | $3.75 → $7.50 from 2027-01-01 | $1.50 |
| Gemini 3.7 Flash | Google (Gemini) | $0.750 → $1.50 from 2027-01-01 | $0.075 → $0.150 from 2027-01-01 | $3.75 → $7.50 from 2027-01-01 | $1.50 |
| Gemini 3.6 Flash | Google (Gemini) | $0.750 → $1.50 from 2027-01-01 | $0.075 → $0.150 from 2027-01-01 | $3.75 → $7.50 from 2027-01-01 | $1.50 |
| GPT-5.4 mini | OpenAI | $0.750 | $0.075 | $4.50 | $1.69 |
| o4-mini | OpenAI | $1.10 | $0.275 | $4.40 | $1.93 |
| Claude Haiku 4.5 | Anthropic (Claude) | $1.00 | $0.100 | $5.00 | $2.00 |
| Gemini 3.5 Flash | Google (Gemini) | $1.50 | $0.150 | $9.00 | $3.38 |
| GPT-5.1 | OpenAI | $1.25 | $0.125 | $10.00 | $3.44 |
| Gemini 2.5 Pro | Google (Gemini) | $1.25 ($2.50 above 200k) | $0.125 ($0.250 above 200k) | $10.00 ($15.00 above 200k) | $3.44 |
| GPT-4.1 | OpenAI | $2.00 | $0.500 | $8.00 | $3.50 |
| o3 | OpenAI | $2.00 | $0.500 | $8.00 | $3.50 |
| Claude Sonnet 5.5 | Anthropic (Claude) | $2.00 | $0.200 | $10.00 | $4.00 |
| Claude Sonnet 5 | Anthropic (Claude) | $2.00 | $0.200 | $10.00 | $4.00 |
| GPT-6.1 Sol | OpenAI | $2.00 | $0.100 | $10.00 | $4.00 |
| GPT-6 Sol | OpenAI | $2.00 | $0.200 | $10.00 | $4.00 |
| GPT-4o | OpenAI | $2.50 | $1.25 | $10.00 | $4.38 |
| GPT-5.6 Terra | OpenAI | $2.00 | $0.200 | $12.00 | $4.50 |
| Gemini 3.1 Pro (Preview) | Google (Gemini) | $2.00 ($4.00 above 200k) | $0.200 ($0.400 above 200k) | $12.00 ($18.00 above 200k) | $4.50 |
| GPT-5.2 | OpenAI | $1.75 | $0.175 | $14.00 | $4.81 |
| GPT-5.4 | OpenAI | $2.50 | $0.250 | $15.00 | $5.63 |
| Claude Sonnet 4.6 | Anthropic (Claude) | $3.00 | $0.300 | $15.00 | $6.00 |
| Claude Sonnet 4.5 | Anthropic (Claude) | $3.00 | $0.300 | $15.00 | $6.00 |
| Claude Opus 5.5 | Anthropic (Claude) | $4.00 | $0.200 | $20.00 | $8.00 |
| GPT-5.6 Sol | OpenAI | $4.00 | $0.400 | $20.00 | $8.00 |
| Claude Opus 5 | Anthropic (Claude) | $5.00 | $0.500 | $25.00 | $10.00 |
| Claude Opus 4.8 | Anthropic (Claude) | $5.00 | $0.500 | $25.00 | $10.00 |
| Claude Opus 4.7 | Anthropic (Claude) | $5.00 | $0.500 | $25.00 | $10.00 |
| Claude Opus 4.6 | Anthropic (Claude) | $5.00 | $0.500 | $25.00 | $10.00 |
| Claude Opus 4.5 | Anthropic (Claude) | $5.00 | $0.500 | $25.00 | $10.00 |
| GPT-5.5 | OpenAI | $5.00 | $0.500 | $30.00 | $11.25 |
| Claude Fable 5.1 | Anthropic (Claude) | $10.00 | $0.250 | $50.00 | $20.00 |
| Claude Fable 5 | Anthropic (Claude) | $10.00 | $1.00 | $50.00 | $20.00 |
| GPT-6 Astra | OpenAI | $10.00 | $1.00 | $50.00 | $20.00 |
| GPT-5.5 Pro | OpenAI | $30.00 | — | $180 | $67.50 |
Why the cheapest is not always the best
- Quality differs a lot. A very cheap model that needs retries or fails your task can cost more than a stronger one. Test on your own data.
- Your mix may differ. If your outputs are long, weight output price more. Use the calculator with your real numbers.
- Caching and batch change the ranking for repeated prompts and offline jobs.
- Prices move. Several providers change prices or run introductory rates; the Gemini Flash models, for example, change price on 2027-01-01.