AI API price index · open-model host · read 16 Oct 2026
Replicate API pricing
25 prices as Replicate publishes them — 1 text models, 5 images and documents, 19 hardware time — read on its own pricing page on 16 Oct 2026. Its current text models cost $10.00 per million output tokens.
Free tier The pricing page states no free tier or credit amount (only a 'Try for free' button). The official billing docs ('Free limits') say: 'You can run select models on Replicate for free, but after a bit you'll be asked to set up billing.'
What to know before you build on Replicate
- Most models bill by the second the hardware runs — an H100 is $0.001525 a second, about $5.49 an hour, unchanged since our 24 September reading.
- Some official models bill per output instead: flux-schnell is $3.00 per thousand images, deepseek-r1 $3.75 in and $10 out per million tokens.
- Multi-GPU tiers and every H200 tier are for committed-spend contracts.
- Private models on dedicated hardware bill for setup and idle time as well as active time.
- Select models can be run free for a while before billing is required; no amount is stated.
Against the other ten
Across the index, 85 current text models run from $0.10 to $50.00 per million output tokens — 500×. Replicate’s cheapest, deepseek-ai / deepseek-r1 at $10.00, has 61 cheaper text models elsewhere. A cheaper model is a different model, not the same one for less — compare capability before price.
Text models
| Model | Input | Cached input | Output | Note |
|---|---|---|---|---|
| deepseek-ai / deepseek-r1official model | $3.75 | — | $10.00 | Output printed per thousand tokens ($0.01 / thousand output tokens), converted to $10.00 per 1M; input printed per million |
Images and documents
| Model | Price | Note |
|---|---|---|
| black-forest-labs / flux-1.1-proofficial model | $0.04 per image | per output image |
| black-forest-labs / flux-devofficial model | $0.025 per image | per output image |
| black-forest-labs / flux-schnellofficial model | $0.003 per image | Printed as $3.00 per thousand output images, converted to $0.003 per image |
| ideogram-ai / ideogram-v3-qualityofficial model | $0.09 per image | per output image |
| recraft-ai / recraft-v3official model | $0.04 per image | per output image |
Hardware time
| Model | Price | Note |
|---|---|---|
| CPU (Small) (cpu-small)hardware | $0.000025 per second | 1x CPU, 2GB RAM; $0.09/hr |
| CPU (cpu)hardware | $0.0001 per second | 4x CPU, 8GB RAM; $0.36/hr |
| Nvidia T4 GPU (gpu-t4)hardware | $0.000225 per second | 1x GPU, 16GB GPU RAM; $0.81/hr |
| Nvidia L40S GPU (gpu-l40s)hardware | $0.000975 per second | 1x GPU, 48GB GPU RAM; $3.51/hr |
| 2x Nvidia L40S GPU (gpu-l40s-2x)hardware | $0.00195 per second | 2x GPU, 96GB GPU RAM; $7.02/hr |
| Nvidia A100 (80GB) GPU (gpu-a100-large)hardware | $0.0014 per second | 1x GPU, 80GB GPU RAM; $5.04/hr |
| 2x Nvidia A100 (80GB) GPU (gpu-a100-large-2x)hardware | $0.0028 per second | 2x GPU, 160GB GPU RAM; $10.08/hr |
| Nvidia H100 GPU (gpu-h100)hardware | $0.001525 per second | 1x GPU, 80GB GPU RAM; $5.49/hr |
| 4x Nvidia A100 (80GB) GPU (gpu-a100-large-4x)hardware (additional) | $0.0056 per second | $20.16/hr; available with committed spend contracts |
| 8x Nvidia A100 (80GB) GPU (gpu-a100-large-8x)hardware (additional) | $0.0112 per second | $40.32/hr; available with committed spend contracts |
| 2x Nvidia H100 GPU (gpu-h100-2x)hardware (additional) | $0.00305 per second | $10.98/hr; available with committed spend contracts |
| 4x Nvidia H100 GPU (gpu-h100-4x)hardware (additional) | $0.0061 per second | $21.96/hr; available with committed spend contracts |
| 8x Nvidia H100 GPU (gpu-h100-8x)hardware (additional) | $0.0122 per second | $43.92/hr; available with committed spend contracts |
| Nvidia H200 GPU (gpu-h200)hardware (additional) | $0.001525 per second | $5.49/hr; 'H200 capacity is available with committed spend contracts.' |
| 2x Nvidia H200 GPU (gpu-h200-2x)hardware (additional) | $0.00305 per second | $10.98/hr; available with committed spend contracts |
| 4x Nvidia H200 GPU (gpu-h200-4x)hardware (additional) | $0.0061 per second | $21.96/hr; available with committed spend contracts |
| 8x Nvidia H200 GPU (gpu-h200-8x)hardware (additional) | $0.0122 per second | $43.92/hr; available with committed spend contracts |
| 4x Nvidia L40S GPU (gpu-l40s-4x)hardware (additional) | $0.0039 per second | $14.04/hr; available with committed spend contracts |
| 8x Nvidia L40S GPU (gpu-l40s-8x)hardware (additional) | $0.0078 per second | $28.08/hr; available with committed spend contracts |
1 of these 25 prices are ranked in the index; the rest say why in their note. Text and embeddings per 1M tokens unless the row says otherwise.
Discounts and surcharges, as published
- Volume discounts: 'We've also got volume discounts for large amounts of spend.' (enterprise)
- Site banner on the pricing page: 'Wan 3.0 is 30% off this week' (no Wan 3.0 price printed on the pricing page)
Where these numbers come from
Read on replicate.com/pricing on 16 Oct 2026, checked against 1 more of Replicate’s own pages. Macrostack measures nothing and is paid by no provider on this index. Prices change; the date is the claim.
Thinking of moving off Replicate? The alternatives, ranked · is Replicate free?
The other providers in the index
- The AI API price index — every provider, cheapest first, and the download
- OpenAI API pricing — model maker
- Anthropic API pricing — model maker
- Google Gemini API pricing — model maker
- Mistral AI API pricing — model maker
- Cohere API pricing — model maker
- Groq API pricing — open-model host
- Together AI API pricing — open-model host
- Fireworks AI API pricing — open-model host
- Deepgram API pricing — voice
- ElevenLabs API pricing — voice
Questions
- How much does the Replicate API cost?
- On 16 Oct 2026, Replicate's current text models cost $10.00 per million output tokens at the standard tier (deepseek-ai / deepseek-r1). The full list, with input and cached-input prices, is on this page.
- Does Replicate have a free tier?
- What Replicate's own pages say: The pricing page states no free tier or credit amount (only a 'Try for free' button). The official billing docs ('Free limits') say: 'You can run select models on Replicate for free, but after a bit you'll be asked to set up billing.'
- Does Replicate offer discounts?
- Volume discounts: 'We've also got volume discounts for large amounts of spend.' (enterprise) Site banner on the pricing page: 'Wan 3.0 is 30% off this week' (no Wan 3.0 price printed on the pricing page)