NotHoly Specter
Backing model, capabilities and customer rates from the published catalog.
Updated October 31, 2026
Base model: GLM 5.3 Abliterated
- Public API ID
notholy-specter- Effective context capacity
- 1,048,576 tokens
- Effective maximum output
- 131,072 tokens
- Model access
- Chat, CLI and supported API contracts
Published capabilities
- Tool calling
- Reasoning
- Streaming
Usage per million tokens
| Token type | Retail USD | Credits |
|---|---|---|
| Input | $3.00 | 300 |
| Cached input | $0.45 | 45 |
| Output | $6.00 | 600 |
USD per 1 million tokens. Schedule retail-2026-10-31-v2. Cached tokens replace ordinary input for that subset. Billable output may include reasoning tokens.
Catalog identity
The NotHoly name is a stable product alias. The catalog identifies its backing model and current limits; NotHoly does not claim to have trained these weights.
Usage and availability
Input, cached input and output have distinct versioned rates. Cached input replaces ordinary input for that subset; it is never counted twice. A published capability belongs to this model, not automatically to every mode or client.