Best-value fast AI models
Fast models are small, fast, cheap models for high-volume and latency-sensitive work. Ranked by price per million output tokens, cheapest first — 14 priced models from the Mintok catalog.
| # | Model | Vendor | Tier | Output $/M | Input $/M | Context |
|---|---|---|---|---|---|---|
| 1 | LLaMA 4 Scout | Meta | 4/8 | $0.30 | $0.10 | 10,000,000 |
| 2 | Gemma 4 26B A4B | 2/8 | $0.33 | $0.06 | 262,144 | |
| 3 | Gemma 4 31B | 3/8 | $0.35 | $0.12 | 262,144 | |
| 4 | Gemini 2.5 Flash-Lite | 3/8 | $0.40 | $0.10 | — | |
| 5 | DeepSeek V4 Flash | DeepSeek | 4/8 | $0.60 | $0.15 | 1,000,000 |
| 6 | Mistral Small 4 | Mistral | 3/8 | $0.60 | $0.15 | 256,000 |
| 7 | GPT-4o mini | OpenAI | 4/8 | $0.60 | $0.15 | 128,000 |
| 8 | GPT-5.6 Luna | OpenAI | 5/8 | $1.20 | $0.20 | 1,050,000 |
| 9 | Magistral Small | Mistral | 3/8 | $1.50 | $0.50 | — |
| 10 | Gemini 3.1 Flash-Lite | 3/8 | $2.50 | $0.30 | — | |
| 11 | Claude Haiku 3.5 | Anthropic | 4/8 | $4 | $0.80 | 200,000 |
| 12 | o4-mini | OpenAI | 3/8 | $4.40 | $1.10 | 200,000 |
| 13 | GPT-5.4 mini | OpenAI | 4/8 | $4.50 | $0.75 | 400,000 |
| 14 | Claude Haiku 4.5 | Anthropic | 5/8 | $5 | $1 | 200,000 |
FAQ
What is the cheapest fast AI model?
The cheapest fast-class model in the Mintok catalog is LLaMA 4 Scout (Meta) at $0.30 per million output tokens.
What is a fast model?
Fast models are small, fast, cheap models for high-volume and latency-sensitive work.
How many fast models does Mintok track?
Mintok tracks 14 priced fast-class models with live $/M-token pricing.
Prices maintained by Mintok's catalog watchers from public sources. Size class + tiers per methodology. Not investment advice.