Pricing
Understand our pricing model for generation and embeddings
Emcie NLP Service uses token-based pricing. Costs vary by model tier and role, with significant savings when using Student models after the automated optimization completes.
Generation Pricing
All prices are per 1 million tokens in USD.
Jackal Tier
| Role | Input Tokens | Output Tokens |
|---|---|---|
| Teacher | $1.00 | $8.00 |
| Student | $0.30 | $2.50 |
Bison Tier
| Role | Input Tokens | Output Tokens |
|---|---|---|
| Teacher | $1.50 | $15.00 |
| Student | $0.90 | $5.00 |
Embedding Pricing
| Model | Input Tokens |
|---|---|
| jackal-embedding | $0.01 |
| bison-embedding | $0.12 |
Token Calculation
Emcie counts tokens using the o200k_base tokenizer, which is available in OpenAI's tiktoken library.
Understanding Automatic Distillation
With automatic distillation, your costs automatically decrease as optimization completes.
Here's an illustrative example (exact numbers may change depending on the use case):
Week 1-2: Teacher pricing (grace period)
Week 3+: Student pricing (optimized)Updated 6 days ago