Models
These are the main public model IDs you should use in clients, SDKs, and editor integrations.
Recommended public model IDs
| Model | Model ID | Max input | Max output | Thinking | Vision | Best for |
|---|---|---|---|---|---|---|
| Claude Opus 4.7 | claude-opus-4.7 | 1M | 32K | Yes | Yes | Hardest reasoning, research, large codebases |
| Claude Opus 4.6 | claude-opus-4.6 | 200K | 32K | Yes | Yes | Complex reasoning, creative writing |
| Claude Opus 4.5 | claude-opus-4.5 | 200K | 32K | Yes | Yes | Complex reasoning, creative writing |
| Claude Sonnet 4.6 | claude-sonnet-4.6 | 1M | 8192 | Yes | Yes | Coding, general tasks, balanced cost |
| Claude Sonnet 4.5 | claude-sonnet-4.5 | 200K | 8192 | Yes | Yes | Coding, everyday development |
| Claude Haiku 4.5 | claude-haiku-4.5 | 200K | 8192 | No | Yes | Fast responses, classification, low-cost automation |
| GPT 5.5 | gpt-5.5 | 200K | 16K | Yes | Yes | Hardest reasoning, research, large codebases |
| GPT 5.4 | gpt-5.4 | 200K | 16K | Yes | Yes | Complex reasoning, creative writing |
Per-model rates
Claudexia uses pay-as-you-go billing — you pay per 1M tokens for each model. Prices are fetched from
| Model | Input / 1M | Output / 1M |
|---|---|---|
| Claude Opus 4.7 | $0.50 | $0.50 |
| Claude Opus 4.6 | $0.50 | $0.50 |
| Claude Opus 4.5 | $0.50 | $0.50 |
| Claude Sonnet 4.6 | $0.33 | $0.33 |
| Claude Sonnet 4.5 | $0.33 | $0.33 |
| Claude Haiku 4.5 | $0.33 | $0.33 |
| GPT 5.5 | $0.50 | $0.50 |
| GPT 5.4 | $0.50 | $0.50 |
How to choose a model
- For coding and development — start with claude-sonnet-4.5. It offers the best balance of speed, quality, and cost for everyday coding tasks.
- For complex reasoning — use claude-opus-4.5 or claude-opus-4.7 when you need deep analysis, multi-step reasoning, or large codebase understanding.
- For fast automation — use claude-haiku-4.5 for classification, extraction, and high-volume tasks where speed and cost matter most.
- For latest capabilities — claude-opus-4.7 and claude-sonnet-4.6 have 1M context window and the newest model improvements.
Notes
- Prefer these stable IDs for new integrations
- Legacy dated Anthropic IDs and some aliases can still resolve, but they are less clear in shared configs
- Use Opus for the heaviest reasoning, Sonnet for general development, and Haiku for fast low-cost work