Model families in the catalog
| Family | IDs in the catalog | Where to start |
|---|---|---|
| Claude Opus | claude-opus-5-5, claude-opus-5, claude-opus-4-8, claude-opus-4-7, claude-opus-4-6 | Architecture, hard debugging, a second review |
| Claude Sonnet | claude-sonnet-5, claude-sonnet-4-6 | Daily coding, support, drafts |
| Claude Fable | claude-fable-5, claude-fable-5-1 | Controlled model comparisons |
| GPT | gpt-5.6-luna, gpt-5.6-terra, gpt-5.6-sol, gpt-6-luna, gpt-6-sol, gpt-6-astra, gpt-6.1-sol | Luna for bulk steps, Sol for code and automation, Astra for hard cases |
| Gemini | gemini-3.1-pro, gemini-3.7-flash, gemini-3.8-flash | Long documents, fast processing |
| DeepSeek | deepseek-v4-pro, deepseek-v4-flash, deepseek-v4.1-flash | High volume, coding |
| Qwen | qwen3.8-flash, qwen3.8-max | Classification, fast text routes |
| GLM | glm-5.2, glm-5.3, glm-5.3-flash | Separate coding and reasoning evaluations |
| Images | gpt-image-2, gpt-image-2-5, nano-banana, nano-banana-2, nano-banana-pro | Style, text legibility, number of manual fixes |
These are catalog IDs, not a promise of a benchmark position, a context size or a price. Your account terms are in the models documentation.
Which model to take for which task
| Task | First route | Second route to compare |
|---|---|---|
| Hard architecture review | claude-opus-5-5 | gpt-6.1-sol |
| Routine code change | claude-sonnet-5 | gpt-6-sol |
| Bulk extraction | gpt-6-luna | deepseek-v4-flash |
| Customer support answer | claude-sonnet-5 | gemini-3.7-flash |
| Long technical document | claude-opus-5-5 | gemini-3.1-pro |
| Image generation | gpt-image-2 | nano-banana-2 |
A route is ready when it passes your fixture set: accuracy, schema compliance, latency, token use, retries and refusals.
Connecting an ID and keeping up with the catalog
Register, create a key and pass the chosen ID to the SDK your project already uses. For an OpenAI-compatible client the base URL looks like this:
from openai import OpenAI
client = OpenAI(
api_key="your_key",
base_url="https://api.codexkey.online/v1",
)
response = client.chat.completions.create(
model="claude-sonnet-5",
messages=[{"role": "user", "content": "Check this answer against the requirements"}],
)
Keep the model ID in configuration and API keys in server-side secrets, and look up protocol details in the API guides. Don't bake the tables above into code. Query GET /v1/models, set an alert for a configured ID that disappears, and keep a fallback on a tested one.
A five-step check on your own tasks
- Collect anonymized requests from the real workflow: closed tickets, fixed bugs, real briefs.
- Write a pass rule for each one.
- Run two IDs with the same prompt and parser.
- Save failures, latency and token counts.
- Move a small slice of traffic before you change the default.
FAQ
Does this list include every live model?
It lists the 32 IDs the public catalog returned on 4 October 2026. Recheck the API and the documentation before deployment, because availability can change.
Where do I find a model's price, and why is it not in this article?
On the pricing page. The catalog gave no stable public price sheet and prices can depend on account terms, so this article records IDs and scenarios as of the update date.
Does one key work for both Claude and GPT?
Yes, if the model is available to your account. Check the ID on the models page or through GET /v1/models.
Which model should be the default?
For a mixed workload, compare Sonnet 5 and GPT 6 Sol on your own fixtures. For more careful reasoning, take Opus 5.5 or GPT 6.1 Sol.
What should I do with an ID that was removed from the catalog?
Keep the old route only during migration, map it to a tested replacement and run the fixtures before you change the default.