Expanded the model lineup — added GPT-5.x and refreshed Claude models

Today we expanded the model catalog in LiteAI. Alongside the familiar Claude models, we added new GPT-5.x entries and refreshed the Claude lineup with a 1M-context version of Opus 4.8. The same LiteAI API key now opens access to the entire list — no extra keys, no separate tariffs, no provider switching.

The full LiteAI model lineup

As of today, eight models are available through a single LiteAI key (sk-bf-…):

Model	Provider	Context	Best for
gpt-5.4	OpenAI	128K	General workhorse: code, text, analysis, reasoning
gpt-5.4-mini	OpenAI	128K	Cheap replacement for simple tasks: classification, summarization, drafts
gpt-5.6-luna	OpenAI	256K	Long context and creative tasks, soft style
gpt-5.6-sol	OpenAI	256K	Precise instructions and technical text, strict format
gpt-5.6-terra	OpenAI	256K	Long reasoning chains and multi-step pipelines
claude-haiku-4-5	Anthropic	200K	Fast replies, low-latency scenarios, chatbots
claude-opus-4-8	Anthropic	200K	Strongest Claude reasoning, code review, architecture
claude-opus-4-8[1m]	Anthropic	1M	Opus with a 1,000,000-token context for big repos and documents
claude-sonnet-4-6	Anthropic	200K	Balance of speed and quality, day-to-day development

All eight models are available in LiteAI Chat (chat.liteai.tech), via the https://api.liteai.tech/v1 endpoint, and inside Claude Code / Cursor / any OpenAI- or Anthropic-compatible client — just swap base_url and paste your LiteAI key.

What's new

New GPT-5.x models. Until today LiteAI offered only Claude. We now added four GPT-5.x entries — gpt-5.4, gpt-5.4-mini, gpt-5.6-luna, gpt-5.6-sol, and gpt-5.6-terra. They aren't a "Claude replacement" — each model shines in its own scenarios:

• gpt-5.4 and gpt-5.4-mini are your everyday workhorses. mini is 5–7× cheaper, so it works great as a first line: preprocessing, request routing, draft generation.
• gpt-5.6-luna, sol, and terra are three specializations of the same line. luna is tuned for creative writing and long context, sol for strict instructions and format, terra for multi-step reasoning and complex pipelines.

Claude Opus 4.8 with 1M context. claude-opus-4-8[1m] is the same Opus 4.8 but with a 1,000,000-token context window. Useful for:

• code review of an entire repo in a single pass;
• analysis of long conversations, logs, regulatory documents;
• tasks that need "memory" of a whole codebase or a book.

The 1M context is billed at the regular input rate — no surcharge for "large context mode".

How to try it

If you already have a LiteAI key, just change the model field in your request. For example, via cURL:

curl https://api.liteai.tech/v1/chat/completions \
  -H "Authorization: Bearer sk-bf-your-key" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-5.4",
    "messages": [{"role": "user", "content": "Hi! Tell me about yourself."}]
  }'

If you don't have a key yet, grab any token package at liteai.tech/pricing: one key works with all eight models, flat 30 ₽ / 1M tokens (no hidden "premium model" surcharge). For a quick test, the smallest 30 ₽ package is enough for a couple dozen Opus 4.8 calls or several hundred Sonnet 4.6 calls.

Inside LiteAI Chat (chat.liteai.tech) you switch models from the dropdown above the input box — the same view shows your current balance and spend.

Bottom line

LiteAI's catalog grew from 4 Claude models to 8: we added GPT-5.4, GPT-5.4-mini, GPT-5.6-luna, GPT-5.6-sol, GPT-5.6-terra, and Claude Opus 4.8 with a 1M context. One key, one tariff, no separate registrations — plug in and try.