Small and fast AI models compared
Cheap, low-latency models for classification, routing, extraction, and cleanup. At these rates per-token cost stops being the constraint and throughput starts being one.
List prices checked
Token prices
| Model | Provider | Input | Output | Cached | Blended | Context |
|---|---|---|---|---|---|---|
| GPT-5 nano | OpenAI | $0.050 | $0.400 | $0.0050 | $0.138 | 400K |
| Gemini 2.5 Flash-Lite | $0.100 | $0.400 | $0.025 | $0.175 | 1M | |
| Mistral Small | Mistral AI | $0.100 | $0.300 | — | $0.150 | 131K |
| GPT-4o mini | OpenAI | $0.150 | $0.600 | $0.075 | $0.262 | 128K |
| Cohere Command R | Cohere | $0.150 | $0.600 | — | $0.262 | 128K |
| GPT-4.1 mini | OpenAI | $0.400 | $1.60 | $0.100 | $0.700 | 1M |
Ranked for: In-app chat assistant
A support or help assistant handling a few thousand conversations a month, with short prompts and short answers.
| # | Model | Input cost | Output cost | Monthly |
|---|---|---|---|---|
| 1 | GPT-5 nano | $1.00 | $2.00 | $3.00 |
| 2 | Mistral Small | $2.00 | $1.50 | $3.50 |
| 3 | Gemini 2.5 Flash-Lite | $2.00 | $2.00 | $4.00 |
| 4 | GPT-4o mini | $3.00 | $3.00 | $6.00 |
| 5 | Cohere Command R | $3.00 | $3.00 | $6.00 |
| 6 | GPT-4.1 mini | $8.00 | $8.00 | $16 |
Ranked for: RAG over documents
Retrieval-augmented answering where each request stuffs several retrieved passages into the prompt, so input dominates.
| # | Model | Input cost | Output cost | Monthly |
|---|---|---|---|---|
| 1 | GPT-5 nano | $15 | $6.00 | $21 |
| 2 | Mistral Small | $30 | $4.50 | $35 |
| 3 | Gemini 2.5 Flash-Lite | $30 | $6.00 | $36 |
| 4 | GPT-4o mini | $45 | $9.00 | $54 |
| 5 | Cohere Command R | $45 | $9.00 | $54 |
| 6 | GPT-4.1 mini | $120 | $24 | $144 |
Ranked for: Coding agent
An agentic loop that reads files and writes patches, generating heavy output and re-sending large context on every turn.
| # | Model | Input cost | Output cost | Monthly |
|---|---|---|---|---|
| 1 | GPT-5 nano | $20 | $40 | $60 |
| 2 | Mistral Small | $40 | $30 | $70 |
| 3 | Gemini 2.5 Flash-Lite | $40 | $40 | $80 |
| 4 | GPT-4o mini | $60 | $60 | $120 |
| 5 | Cohere Command R | $60 | $60 | $120 |
| 6 | GPT-4.1 mini | $160 | $160 | $320 |