OpenLuxAnthropic models API
Claude Opus 5, Claude Sonnet 4.5 and Claude Fable 5.1 are available on OpenLux through the same OpenAI-compatible endpoint as every other model, at 82-91% below official list rates.
Calling Anthropic models on OpenLux
Anthropic models are known for long-context work, careful instruction following and strong coding ability. Claude Opus 5 is the flagship for the hardest reasoning tasks, Claude Sonnet 4.5 is the balanced daily driver for coding and tool use, and Claude Fable 5.1 leans toward long-form creation and research where output quality compounds over thousands of words.
On OpenLux all three speak the OpenAI chat-completions protocol, so a team already using the OpenAI SDK can reach Claude by changing the model name rather than the client. That makes A/B testing straightforward: send the same prompt to Sonnet and to a GPT tier, keep whichever answers better.
Rates are metered per token in USD — $0.4412 per 1M tokens for Opus 5, $0.2647 for Sonnet 4.5 and $1.765 for Fable 5.1 — with the console showing cost per request while the workload runs.
Why teams call Anthropic on OpenLux
Long context
Comfortable with large inputs, which suits repository-scale code review and long document work.
Coding quality
Sonnet 4.5 balances speed, correctness and price for everyday engineering tasks.
Careful instruction following
Holds tone, structure and constraints across long outputs where drift is the usual failure mode.
Tool use
Function calling and vision input are available on the same endpoint, so agents need no special client.
Rates for Anthropic models
| Model | Category | Rate | Discount |
|---|---|---|---|
| claude-opus-5 | Text | $0.4412 per 1M tokens | -91% |
| claude-sonnet-4-5-20250929 | Text | $0.2647 per 1M tokens | -91% |
| claude-fable-5-1 | Text | $1.765 per 1M tokens | -82% |
Rates are USD, metered pay as you go with no subscription. See the full model catalogue and pricing plans for credit allowances.
Where Anthropic models fit
Software engineering
Code review, refactoring and agentic editing in editors and terminal agents that expect an OpenAI-style API.
Long-form writing
Reports, documentation and marketing copy that must stay consistent across many sections.
Research synthesis
Read long sources, extract the argument and produce a structured summary with citations.
Anthropic on OpenLux questions
How much do Claude models cost on OpenLux?
Claude Opus 5 is $0.4412 per 1M tokens (91% below list), Claude Sonnet 4.5 is $0.2647 (91% below list) and Claude Fable 5.1 is $1.765 (82% below list), all metered pay as you go.
Can I use the Anthropic SDK instead?
You can use any client that speaks the OpenAI protocol through OpenLux, which includes the official OpenAI SDKs and the many tools built on them.
Which Claude model should I start with?
Sonnet 4.5 covers most production traffic; move to Opus 5 when reasoning depth is the bottleneck and to Fable 5.1 for long-form output that must stay polished.
Do Claude models support tools and images?
Yes. Tool calling, vision input and streamed responses run on the same endpoint and the same key as the rest of the catalogue.
Start calling Anthropic models today
One OpenLux key covers Anthropic and every other provider in the catalogue, metered in USD with real-time cost in the console.