Overview
Credits System
Users can purchase API credits with crypto. These credits are used to access Heurist’s decentralized AI compute resources, with different API endpoints consuming credits based on their computational requirements.
- 1 USD = 100 credits
- When paying with HEU tokens, users receive a 20% bonus (120 credits per $1 equivalent)
Credits can be purchased at heurist.ai/credits.
Credits Consumption
1. Image Generation
Base pricing is calculated for 1024x1024 resolution with 20 iterations. Other resolutions or iteration counts are priced proportionally.
| Model | Credits per Image |
|---|---|
| Stable Diffusion XL/1.5 | 0.05 |
| Stable Diffusion with LoRA | 0.1 |
| FLUX.1-dev | 0.2 |
| FLUX.1-dev with LoRA | 0.5 |
2. Video Generation
| Model | Duration | Credits |
|---|---|---|
| Hunyuan | 5 seconds | 25 |
3. Large Language Models
Pricing is in credits per million tokens. Each model has separate pricing for input tokens (what you send as prompts) and output tokens (what the model generates).
| Model | Input (per 1M tokens) | Output (per 1M tokens) |
|---|---|---|
| nvidia/llama-3.1-nemotron-70b-instruct | 15 | 30 |
| qwen/qwen-2.5-coder-32b-instruct | 10 | 18 |
| mistralai/mixtral-8x7b-instruct | 25 | 25 |
| mistralai/mixtral-8x22b-instruct | 100 | 100 |
| meta-llama/llama-3.3-70b-instruct | 15 | 30 |
| NousResearch/Hermes-3-Llama-3.1-8B | 10 | 10 |
| deepseek/deepseek-r1 | 300 | 300 |
| deepseek/deepseek-v3 | 100 | 100 |
| deepseek/deepseek-r1-distill-llama-70b | 80 | 80 |
| NaniDAO/deepseek-r1-qwen-2.5-32B-ablated | 30 | 30 |
| mistralai/mistral-small-24b-instruct | 30 | 30 |
For example, if you send a prompt of 1,000 tokens to Mixtral 8x7B and receive a 500-token response:
- Input cost: (1,000 / 1,000,000) × 25 = 0.025 credits
- Output cost: (500 / 1,000,000) × 25 = 0.0125 credits
- Total cost: 0.0375 credits
4. Embedding Models
Pricing is in credits per million tokens.
| Model | Input (per 1M tokens) |
|---|---|
| baaiai/bge-large-en-v1.5 | 1 |