FREE
$0forever
10,000 credits / mo · ~$10 value
For experimentation and prototyping.
- ✓ 10 requests / min
- ✓ 1 API key
- ✓ All 30+ models
- ✓ Community support
- ✓ No credit card required
◦ PRICING
Credit-based pricing.
1 credit = $0.001. Pay for what you use. Small models cost less, large models cost more.
PROPOPULAR
$49/ month
500,000 credits / mo · ~$500 value
For production applications.
- ✓ 100 requests / min
- ✓ 5 API keys
- ✓ Overage $1.00 / 1K credits
- ✓ Email support
- ✓ 30-day audit logs
- ✓ Usage alerts + per-key breakdown
ENTERPRISEDEDICATED
Custom
Unlimited credits · negotiated rates
For regulated industries.
- ✓ Custom rate limits
- ✓ Unlimited API keys
- ✓ Volume discount on credits
- ✓ Priority model access
- ✓ Dedicated support + SLA
- ✓ 1-year audit logs + export
- ✓ SOX / PIPEDA reporting
- ✓ Dedicated model hosting
Credits per model
Smaller models = fewer credits. Pick the right model for your use case.
Chat · Reasoning · Code
| MODEL | CATEGORY | CREDITS / 1M IN | CREDITS / 1M OUT |
|---|---|---|---|
| GPT-OSS 20B | LLM | 60 | 230 |
| Qwen 3 Coder 30B | Code | 90 | 330 |
| GPT-OSS 120B | LLM | 110 | 590 |
| Qwen 3 32B | LLM | 110 | 310 |
| Mistral Small 3.2 24B | LLM | 130 | 390 |
| Llama 3.1 8B Instruct | LLM | 140 | 140 |
| Mistral 7B Instruct v0.3 | LLM | 140 | 140 |
| Mistral Nemo 12B | LLM | 180 | 180 |
| Mixtral 8x7B Instruct | LLM | 880 | 880 |
| Llama 3.3 70B Instruct | LLM | 930 | 930 |
| DeepSeek R1 Distill 70B | Reasoning | 930 | 930 |
Embedding
| MODEL | CREDITS / 1M TOKENS |
|---|---|
| BGE-M3 | 10 |
| BGE Base EN v1.5 | 10 |
| BGE Multilingual Gemma2 | 10 |
| Qwen 3 Embedding 8B | 150 |
Vision · Speech · Image · Safety
| MODEL | CATEGORY | CREDITS |
|---|---|---|
| Qwen 2.5 VL 72B | Vision | 1,260 / M tokens |
| Whisper Large V3 | Speech | Included |
| Whisper Large V3 Turbo | Speech | Included |
| Stable Diffusion XL Base | Image | Included |
| Qwen 3 Guard 8B | Safety | Included |
| Qwen 3 Guard 0.6B | Safety | Included |
HOW CREDITS WORK
- 01
Each API call costs credits based on the model and tokens used.
- 02
Small models (Llama 8B, Mistral 7B) cost ~140 credits per million tokens.
- 03
Large models (GPT-OSS 120B, Llama 70B) cost ~900 credits per million tokens.
- 04
Embedding and safety models are ultra-cheap or included at no additional cost.
- 05
Your dashboard shows real-time credit balance, burn rate, and per-key breakdown.