AI cost calculator
Estimate the monthly API cost of a workload across current models and providers. Below: three preset workloads priced live at each model’s standard headline price.
For a support chatbot with 50,000 requests a month (1,500 input and 400 output tokens each), standard API prices range from $10.35 a month with DeepSeek V4.1 Flash to $1,750.00 with GPT-6 Astra Pro. Enter your own workload in the calculator.
Support chatbot: 50,000 requests a month
Short multi-turn conversations with a system prompt and recent history. 1,500 input and 400 output tokens per request. Open this preset in the calculator
| Model | Monthly cost (standard) | Batch tier, if offered | Input / 1M | Output / 1M | Provider |
|---|---|---|---|---|---|
| DeepSeek V4.1 Flash | $10.35 | $6.75 | $0.09 | $0.18 | Decartfp4 |
| GLM 5.3 Flash | $10.63 | $7.35 | $0.075 | $0.25 | DeepInfrafp4 |
| Claude Haiku 5.5 | $17.50 | $8.75 | $0.10 | $0.50 | Anthropic |
| GPT-6 Luna | $17.50 | $8.75 | $0.10 | $0.50 | OpenAI |
| DeepSeek V4 Pro 0813 | $18.00 | Not offered | $0.1333 | $0.40 | Baidufp8 |
| MiniMax M3 | $36.45 | Not offered | $0.23 | $0.96 | CoreWeavefp4 |
| MiMo-V2.6-Pro | $49.65 | Not offered | $0.43 | $0.87 | DeepInfrafp8 |
| GLM 5.3 | $67.60 | No cheaper batch price | $0.50 | $1.51 | Sail Researchfp8 |
| Gemini 3.5 Flash Lite | $72.50 | $36.25 | $0.30 | $2.50 | Google AI Studio |
| Mistral Large 4 | $92.80 | Not offered | $0.68 | $2.09 | Mistral AI |
| Step 5 Preview | $129.00 | Not offered | $1.00 | $2.70 | StepFunfp8 |
| Gemini 3.8 Flash | $131.25 | $65.63 | $0.75 | $3.75 | Google AI Studio |
| Muse Spark 1.3 | $178.75 | Not offered | $1.25 | $4.25 | Meta |
| Kimi K3 | $199.80 | No cheaper batch price | $0.504 | $8.10 | Sail Researchfp4 |
| Grok 4.7 | $270.00 | Not offered | $2.00 | $6.00 | xAI |
The 15 cheapest of 24 tracked flagship and small models, cheapest first.
RAG / research assistant: 20,000 requests a month
Retrieved passages dominate the prompt; answers cite and summarise them. 8,000 input and 700 output tokens per request. Open this preset in the calculator
| Model | Monthly cost (standard) | Batch tier, if offered | Input / 1M | Output / 1M | Provider |
|---|---|---|---|---|---|
| GLM 5.3 Flash | $15.50 | $10.52 | $0.075 | $0.25 | DeepInfrafp4 |
| DeepSeek V4.1 Flash | $16.92 | $10.10 | $0.09 | $0.18 | Decartfp4 |
| Claude Haiku 5.5 | $23.00 | $11.50 | $0.10 | $0.50 | Anthropic |
| GPT-6 Luna | $23.00 | $11.50 | $0.10 | $0.50 | OpenAI |
| DeepSeek V4 Pro 0813 | $26.93 | Not offered | $0.1333 | $0.40 | Baidufp8 |
| MiniMax M3 | $50.24 | Not offered | $0.23 | $0.96 | CoreWeavefp4 |
| MiMo-V2.6-Pro | $80.98 | Not offered | $0.43 | $0.87 | DeepInfrafp8 |
| Gemini 3.5 Flash Lite | $83.00 | $41.50 | $0.30 | $2.50 | Google AI Studio |
| GLM 5.3 | $101.07 | $100.00 | $0.50 | $1.51 | Sail Researchfp8 |
| Mistral Large 4 | $138.06 | Not offered | $0.68 | $2.09 | Mistral AI |
| Gemini 3.8 Flash | $172.50 | $86.25 | $0.75 | $3.75 | Google AI Studio |
| Kimi K3 | $194.04 | No cheaper batch price | $0.504 | $8.10 | Sail Researchfp4 |
| Step 5 Preview | $197.80 | Not offered | $1.00 | $2.70 | StepFunfp8 |
| Muse Spark 1.3 | $259.50 | Not offered | $1.25 | $4.25 | Meta |
| Grok 4.7 | $404.00 | Not offered | $2.00 | $6.00 | xAI |
The 15 cheapest of 24 tracked flagship and small models, cheapest first.
Batch summarisation: 10,000 requests a month
Long documents summarised offline in bulk. 10,000 input and 500 output tokens per request. Open this preset in the calculator
| Model | Monthly cost (standard) | Batch tier, if offered | Input / 1M | Output / 1M | Provider |
|---|---|---|---|---|---|
| GLM 5.3 Flash | $8.75 | $5.90 | $0.075 | $0.25 | DeepInfrafp4 |
| DeepSeek V4.1 Flash | $9.90 | $5.75 | $0.09 | $0.18 | Decartfp4 |
| Claude Haiku 5.5 | $12.50 | $6.25 | $0.10 | $0.50 | Anthropic |
| GPT-6 Luna | $12.50 | $6.25 | $0.10 | $0.50 | OpenAI |
| DeepSeek V4 Pro 0813 | $15.33 | Not offered | $0.1333 | $0.40 | Baidufp8 |
| MiniMax M3 | $27.80 | Not offered | $0.23 | $0.96 | CoreWeavefp4 |
| Gemini 3.5 Flash Lite | $42.50 | $21.25 | $0.30 | $2.50 | Google AI Studio |
| MiMo-V2.6-Pro | $47.35 | Not offered | $0.43 | $0.87 | DeepInfrafp8 |
| GLM 5.3 | $57.53 | $55.00 | $0.50 | $1.51 | Sail Researchfp8 |
| Mistral Large 4 | $78.45 | Not offered | $0.68 | $2.09 | Mistral AI |
| Kimi K3 | $90.90 | No cheaper batch price | $0.504 | $8.10 | Sail Researchfp4 |
| Gemini 3.8 Flash | $93.75 | $46.88 | $0.75 | $3.75 | Google AI Studio |
| Step 5 Preview | $113.50 | Not offered | $1.00 | $2.70 | StepFunfp8 |
| Muse Spark 1.3 | $146.25 | Not offered | $1.25 | $4.25 | Meta |
| Grok 4.7 | $230.00 | Not offered | $2.00 | $6.00 | xAI |
The 15 cheapest of 24 tracked flagship and small models, cheapest first.
How the estimate works
Monthly cost = requests × (input tokens × input price + output tokens × output price) ÷ 1,000,000, at each model’s standard headline price (cheapest qualifying default-region provider). Caching, long-context surcharges, reasoning tokens beyond the stated output and taxes are not included. Batch prices come from the provider’s batch tier or batch endpoint when one is listed. Estimates are indicative, not quotes.