LiteAI adds Claude Sonnet 5 and Claude Opus 5

LiteAI has added two new Anthropic models to its catalog: claude-sonnet-5 and claude-opus-5. Both are available through the same LiteAI API key as the existing Claude models.

New models

Model	Best for
claude-sonnet-5	Everyday development, analysis, writing, and tasks that need a balance of quality and speed
claude-opus-5	Complex reasoning, architecture, deep code review, and tasks where maximum quality matters most

Sonnet 5 is the practical default for most workloads. Choose Opus 5 when deeper analysis and result quality matter more than the lowest latency.

How to connect

If you already have a LiteAI key, set the new model in the model field.

For the OpenAI-compatible API:

curl https://api.liteai.tech/v1/chat/completions \
  -H "Authorization: Bearer sk-bf-your-key" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "claude-sonnet-5",
    "messages": [{"role": "user", "content": "Hello!"}],
    "max_tokens": 1024
  }'

For the Anthropic Messages API, use the same model names with https://api.liteai.tech/anthropic/v1/messages.

Where the models are available

You can use the new models:

• through the LiteAI API;
• in Claude Code and Claude Desktop;
• in Cursor, Zed, opencode, and other compatible clients;
• in LiteAI Chat when it is connected to your key.

One key works with both claude-sonnet-5 and claude-opus-5; no separate registration is required. See the LiteAI documentation for the full model list and configuration examples.

Which model should you choose?

Start with claude-sonnet-5 for everyday tasks, coding, document analysis, and quick iterations. Switch to claude-opus-5 when you need complex decomposition, architecture decisions, or especially careful review. A complete “task → model → price” matrix (covering Sonnet 5, Opus 5, and fable-5) is in a separate article “Choosing a Claude model in 2026: a table by task and budget”.

Try the new models with the same LiteAI token packages — see the pricing page for details. All of the extra models for 2026 (GPT-5.x, the Claude refresh, the new releases) — in a separate article “New models in 2026”. And the flagship fable-5 (long context, reasoning) — in “The fable-5 model — what's new”.