Balanced AI models compared
Models that trade a slice of frontier capability for a large fraction of the price. This is where most production traffic belongs, and where the price-per-quality curve is steepest.
List prices checked
Token prices
| Model | Provider | Input | Output | Cached | Blended | Context |
|---|---|---|---|---|---|---|
| DeepSeek-V4-Flash | DeepSeek | $0.140 | $0.280 | $0.0028 | $0.175 | 1M |
| GPT-5 mini | OpenAI | $0.250 | $2.00 | $0.025 | $0.688 | 400K |
| Gemini 2.5 Flash | $0.300 | $2.50 | $0.075 | $0.850 | 1M | |
| Grok 3 mini | xAI | $0.300 | $0.500 | — | $0.350 | 131K |
| Codestral | Mistral AI | $0.300 | $0.900 | — | $0.450 | 262K |
| Claude Haiku 4.5 | Anthropic | $1.00 | $5.00 | $0.100 | $2.00 | 200K |
| OpenAI o4-mini | OpenAI | $1.10 | $4.40 | $0.275 | $1.93 | 200K |
| GPT-4.1 | OpenAI | $2.00 | $8.00 | $0.500 | $3.50 | 1M |
| GPT-4o | OpenAI | $2.50 | $10.00 | $1.25 | $4.38 | 128K |
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 | DeepSeek-V4-Flash | $2.80 | $1.40 | $4.20 |
| 2 | Grok 3 mini | $6.00 | $2.50 | $8.50 |
| 3 | Codestral | $6.00 | $4.50 | $11 |
| 4 | GPT-5 mini | $5.00 | $10 | $15 |
| 5 | Gemini 2.5 Flash | $6.00 | $13 | $19 |
| 6 | OpenAI o4-mini | $22 | $22 | $44 |
| 7 | Claude Haiku 4.5 | $20 | $25 | $45 |
| 8 | GPT-4.1 | $40 | $40 | $80 |
| 9 | GPT-4o | $50 | $50 | $100 |
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 | DeepSeek-V4-Flash | $42 | $4.20 | $46 |
| 2 | Grok 3 mini | $90 | $7.50 | $98 |
| 3 | Codestral | $90 | $14 | $104 |
| 4 | GPT-5 mini | $75 | $30 | $105 |
| 5 | Gemini 2.5 Flash | $90 | $38 | $128 |
| 6 | Claude Haiku 4.5 | $300 | $75 | $375 |
| 7 | OpenAI o4-mini | $330 | $66 | $396 |
| 8 | GPT-4.1 | $600 | $120 | $720 |
| 9 | GPT-4o | $750 | $150 | $900 |
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 | DeepSeek-V4-Flash | $56 | $28 | $84 |
| 2 | Grok 3 mini | $120 | $50 | $170 |
| 3 | Codestral | $120 | $90 | $210 |
| 4 | GPT-5 mini | $100 | $200 | $300 |
| 5 | Gemini 2.5 Flash | $120 | $250 | $370 |
| 6 | OpenAI o4-mini | $440 | $440 | $880 |
| 7 | Claude Haiku 4.5 | $400 | $500 | $900 |
| 8 | GPT-4.1 | $800 | $800 | $1,600 |
| 9 | GPT-4o | $1,000 | $1,000 | $2,000 |