The Mintok catalogue
Every lab ships the same ladder.
Flagship, balanced, fast, nano — each provider sells the same four rungs. Capability tier tells you how smart a model is; size classtells you which rung it sits on. They don’t line up across vendors — which is why comparing on tier alone misleads. Priced in $ per million output tokens.
Loading the current lineup…
Which of these is cheapest for your prompts?
List prices don't answer that — measured cost, speed, and judged quality on your actual prompts do. A benchmark run covers the leading models and highlights the cheapest one that clears your quality bar.
All models (46)
Every model in the catalog — live pricing, specs and $/M-token on each page.
- Qwen 2.5 72B$0.4/M
- Qwen 3 235B (MoE)$1.82/M
- Qwen 3 30B (MoE)$0.5/M
- Qwen2.5 3BT1
- Claude Opus 4$75/M
- Claude Opus 4.7T7
- Claude Opus 4.8$25/M
- Claude Sonnet 4$15/M
- Claude Sonnet 4.6$15/M
- Claude Haiku 3.5$4/M
- Command R+$10/M
- DeepSeek R1$2.5/M
- DeepSeek V3$0.8001/M
- DeepSeek-V3 (MoE)T5
- Gemini 2.5 Pro$10/M
- Gemini 2.5 Flash$2.5/M
- Gemma 3 27B$0.33/M
- Gemma 3 12BT2
- Gemma 2 2BT1
- Gemma 3 1BT1
- Gemma 3 4BT1