Cheapest LLM API Models
Compare the lowest-cost current LLM API models by input price, output price, context window, and example total cost.
Quick answer
Cheapest current overall
Command R7B
Cheapest OpenAI
GPT-5 nano
Cheapest Claude
Claude Haiku 4.5
Cheapest Gemini
Gemini 2.5 Flash-Lite
Cheapest long-context
Gemini 2.5 Flash-Lite
Current models ranked by 1M input + 1M output cost
| Rank | Provider | Model | Input / 1M | Output / 1M | Context window | Total (1M + 1M) | Status |
|---|---|---|---|---|---|---|---|
| 1 | Cohere | Command R7B | $0.0375 | $0.1500 | 128,000 | $0.1875 | CURRENT |
| 2 | Mistral | Ministral 3 3B | $0.1000 | $0.1000 | 256,000 | $0.2000 | CURRENT |
| 3 | Mistral | Ministral 3 8B | $0.1500 | $0.1500 | 256,000 | $0.3000 | CURRENT |
| 4 | Mistral | Ministral 14 8B | $0.2000 | $0.2000 | 256,000 | $0.4000 | CURRENT |
| 5 | DeepSeek | DeepSeek V4 Flash | $0.1400 | $0.2800 | 1,000,000 | $0.4200 | CURRENT |
| 6 | OpenAI | GPT-5 nano | $0.0500 | $0.4000 | 400,000 | $0.4500 | CURRENT |
| 7 | Gemini 2.5 Flash-Lite | $0.1000 | $0.4000 | 1,048,576 | $0.5000 | CURRENT | |
| 8 | OpenAI | GPT-4.1 nano | $0.1000 | $0.4000 | 1,047,576 | $0.5000 | CURRENT |
| 9 | Cohere | Command R | $0.1500 | $0.6000 | 128,000 | $0.7500 | CURRENT |
| 10 | Mistral | Mistral Small 4 | $0.1500 | $0.6000 | 256,000 | $0.7500 | CURRENT |
| 11 | OpenAI | GPT-4o mini | $0.1500 | $0.6000 | 128,000 | $0.7500 | CURRENT |
| 12 | OpenAI | GPT-5.4 nano | $0.2000 | $1.2500 | 400,000 | $1.4500 | CURRENT |
| 13 | Gemini 3.1 Flash-Lite | $0.2500 | $1.5000 | 1,048,576 | $1.7500 | CURRENT | |
| 14 | Mistral | Mistral Large 3 | $0.5000 | $1.5000 | 256,000 | $2.0000 | CURRENT |
| 15 | OpenAI | GPT-4.1 mini | $0.4000 | $1.6000 | 1,047,576 | $2.0000 | CURRENT |
| 16 | OpenAI | GPT-5 mini | $0.2500 | $2.0000 | 400,000 | $2.2500 | CURRENT |
| 17 | Mistral | Mistral Medium 3.1 | $0.4000 | $2.0000 | 128,000 | $2.4000 | CURRENT |
| 18 | Gemini 2.5 Flash | $0.3000 | $2.5000 | 1,048,576 | $2.8000 | CURRENT | |
| 19 | xAI | Grok 4.3 | $1.2500 | $2.5000 | 1,000,000 | $3.7500 | CURRENT |
| 20 | DeepSeek | DeepSeek V4 Pro | $1.7400 | $3.4800 | 1,000,000 | $5.2200 | CURRENT |
| 21 | OpenAI | GPT-5.4 mini | $0.7500 | $4.5000 | 400,000 | $5.2500 | CURRENT |
| 22 | OpenAI | o1-mini | $1.1000 | $4.4000 | 128,000 | $5.5000 | CURRENT |
| 23 | OpenAI | o3-mini | $1.1000 | $4.4000 | 200,000 | $5.5000 | CURRENT |
| 24 | OpenAI | o4-mini | $1.1000 | $4.4000 | 200,000 | $5.5000 | CURRENT |
| 25 | Anthropic | Claude Haiku 4.5 | $1.0000 | $5.0000 | 200,000 | $6.0000 | CURRENT |
| 26 | Mistral | Magistral Medium | $2.0000 | $5.0000 | 128,000 | $7.0000 | CURRENT |
| 27 | Mistral | Mistral Medium 3.5 | $1.5000 | $7.5000 | 256,000 | $9.0000 | CURRENT |
| 28 | OpenAI | GPT-4.1 | $2.0000 | $8.0000 | 1,047,576 | $10.0000 | CURRENT |
| 29 | OpenAI | o3 | $2.0000 | $8.0000 | 200,000 | $10.0000 | CURRENT |
| 30 | Gemini 3.5 Flash | $1.5000 | $9.0000 | 1,048,576 | $10.5000 | CURRENT | |
| 31 | Gemini 2.5 Pro | $1.2500 | $10.0000 | 1,048,576 | $11.2500 | CURRENT | |
| 32 | OpenAI | GPT-5 | $1.2500 | $10.0000 | 400,000 | $11.2500 | CURRENT |
| 33 | OpenAI | GPT-5.1 | $1.2500 | $10.0000 | 400,000 | $11.2500 | CURRENT |
| 34 | Cohere | Command A | $2.5000 | $10.0000 | 256,000 | $12.5000 | CURRENT |
| 35 | Cohere | Command R+ | $2.5000 | $10.0000 | 128,000 | $12.5000 | CURRENT |
| 36 | OpenAI | GPT-4o | $2.5000 | $10.0000 | 128,000 | $12.5000 | CURRENT |
| 37 | Gemini 3.1 Pro | $2.0000 | $12.0000 | 1,048,576 | $14.0000 | CURRENT | |
| 38 | OpenAI | GPT-5.2 | $1.7500 | $14.0000 | 400,000 | $15.7500 | CURRENT |
| 39 | OpenAI | GPT-5.4 | $2.5000 | $15.0000 | 1,050,000 | $17.5000 | CURRENT |
| 40 | Anthropic | Claude Sonnet 4.5 | $3.0000 | $15.0000 | 1,000,000 | $18.0000 | CURRENT |
| 41 | Anthropic | Claude Sonnet 4.6 | $3.0000 | $15.0000 | 1,000,000 | $18.0000 | CURRENT |
| 42 | Anthropic | Claude Opus 4.5 | $5.0000 | $25.0000 | 200,000 | $30.0000 | CURRENT |
| 43 | Anthropic | Claude Opus 4.6 | $5.0000 | $25.0000 | 1,000,000 | $30.0000 | CURRENT |
| 44 | Anthropic | Claude Opus 4.7 | $5.0000 | $25.0000 | 1,000,000 | $30.0000 | CURRENT |
| 45 | OpenAI | GPT-5.5 | $5.0000 | $30.0000 | 1,050,000 | $35.0000 | CURRENT |
| 46 | OpenAI | o1 | $15.0000 | $60.0000 | 200,000 | $75.0000 | CURRENT |
| 47 | Anthropic | Claude Opus 4.1 | $15.0000 | $75.0000 | 200,000 | $90.0000 | CURRENT |
| 48 | OpenAI | o3-pro | $20.0000 | $80.0000 | 200,000 | $100.0000 | CURRENT |
| 49 | OpenAI | GPT-5 pro | $15.0000 | $120.0000 | 400,000 | $135.0000 | CURRENT |
| 50 | OpenAI | GPT-5.2 pro | $21.0000 | $168.0000 | 400,000 | $189.0000 | CURRENT |
| 51 | OpenAI | GPT-5.4 pro | $30.0000 | $180.0000 | 1,050,000 | $210.0000 | CURRENT |
| 52 | OpenAI | GPT-5.5 pro | $30.0000 | $180.0000 | 1,050,000 | $210.0000 | CURRENT |
| 53 | OpenAI | o1-pro | $150.0000 | $600.0000 | 200,000 | $750.0000 | CURRENT |
Cheapest input token prices
- Cohere · Command R7B $0.0375
- OpenAI · GPT-5 nano $0.0500
- Mistral · Ministral 3 3B $0.1000
- Google · Gemini 2.5 Flash-Lite $0.1000
- OpenAI · GPT-4.1 nano $0.1000
Cheapest output token prices
- Mistral · Ministral 3 3B $0.1000
- Cohere · Command R7B $0.1500
- Mistral · Ministral 3 8B $0.1500
- Mistral · Ministral 14 8B $0.2000
- DeepSeek · DeepSeek V4 Flash $0.2800
Cheapest large-context models
- Cohere · Command R7B 128K
- Mistral · Ministral 3 3B 256K
- Mistral · Ministral 3 8B 256K
- Mistral · Ministral 14 8B 256K
- DeepSeek · DeepSeek V4 Flash 1M
Why the cheapest model is not always the best choice
Price is one variable. Context needs, output quality requirements, latency targets, and workload complexity can make a slightly higher-priced model more cost-effective overall.
Compare with your own token usage
Open the calculator and enter your real input and output token assumptions.
Open calculatorFAQ
What is the cheapest LLM API model?
It changes over time. This page ranks current models by measurable token-cost data.
Are cheap LLM API models good enough for production?
Sometimes. Validate quality, latency, and reliability for your specific workload.
Why can output tokens change total cost so much?
Many providers price output tokens higher, so long responses can dominate cost.
Which provider has the cheapest current models?
Use the ranked table to compare providers directly under the same cost formula.
How can I reduce LLM API costs?
Trim prompts, constrain output length, and benchmark lower-cost models first.
Hosting your AI app?
After comparing API costs, the next cost factor is where your app runs. DigitalOcean can be a simple option for hosting prototypes, API backends, workers, databases, and Laravel apps.
Explore DigitalOcean →This link is an affiliate link. This means that, at zero cost to you, we earn commissions when you shop through the link.