Claude Opus 5
Claude Opus 5 by Anthropic: $5 input and $25 output per 1M tokens, 1M context. Call it through the odnoga LLM gateway.
Claude Opus 5 is served by Anthropic and called through odnoga with the same OpenAI-compatible request shape as every other model in the catalog. It accepts images alongside text. It supports tool and function calling. Its 1M-token context window sets how much input you can send in one request. Cached input is billed at $0.5 per 1M tokens, so repeated prefixes cost less.
Specification and price
| Vendor | Anthropic |
|---|---|
| Model ID | claude-opus-5 |
| Context window | 1M tokens |
| Max output | 128K tokens |
| Capabilities | Vision, Tools, Streaming |
| Per 1M tokens | Vendor cost | Your price on Free+7% |
|---|---|---|
| Input | $5 | $5.38 |
| Output | $25 | $26.88 |
| Cached input | $0.5 | $0.538 |
Vendor cost is the list price per million tokens as recorded in the odnoga catalog; your price applies your plan margin with the same formula that bills every request — see pricing. Pricing
Call it through odnoga
const res = await fetch("https://api.odnoga.com/v1/chat/completions", {
method: "POST",
headers: {
Authorization: `Bearer ${process.env.ODNOGA_API_KEY}`,
"Content-Type": "application/json",
},
body: JSON.stringify({
model: "claude-opus-5",
messages: [{ role: "user", content: "Hello" }],
}),
});
Same request shape for every vendor in the catalog — swap the model id and odnoga handles keys, routing, limits and cost accounting.
How to use Claude Opus 5
Reviewed 2026-09-14
Best for
- Complex agentic coding and enterprise knowledge work — the use Anthropic states for this model.
- Long-horizon tasks: multi-hour agent runs, migrations, large refactors, deep document review.
- Work over very large inputs, with a 1M-token context window and 128K max output.
- Careful drafting where quality of the written answer matters more than price per call.
Practical tips through odnoga
- 01Thinking is on by default in this generation and disabling it is restricted — plan for the extra output tokens instead of trying to switch them off.
- 02Structure the prompt so the long, stable part comes first; prompt caching then cuts the cost of repeated context sharply.
- 03Give agents explicit stop conditions; long-horizon capability means long-horizon spend if the loop has no exit.
- 04Compare Opus 5 against Sonnet 5 on the same frozen test cases before making it the default — the gap is task-specific.
What a month costs
1,000 calls a month, 10,000 input tokens and 2,000 output tokens each, at your Free price (vendor cost +7%):
| Input (10M tokens) | $53.76 |
|---|---|
| Output (2M tokens) | $53.76 |
| Your cost per month on Free | $108 |
Vendor list cost $100 + odnoga margin $7.53 (+7%). Cached input or answer reuse lowers it; odnoga records both numbers per request.
Claude Opus 5 compared
| Model | Context | Input / 1M | Output / 1M | Capabilities |
|---|---|---|---|---|
| Claude Opus 5 | 1M | $5 | $25 | Vision, Tools, Streaming |
| Claude Fable 5 | 1M | $10 | $50 | Vision, Tools, Streaming |
| Claude Fable 5.1 | 1M | $10 | $50 | Vision, Tools, Streaming |
Questions
- What is Claude Opus 5 best at?
- Anthropic positions it for complex agentic coding and enterprise work, describing gains over Opus 4.8 in deep reasoning, agentic and long-horizon tasks.
- Opus 5 or Sonnet 5?
- Sonnet 5 is the speed-and-intelligence balance point; Opus 5 is the heavier option for the hardest work. Run both on your own test cases in the odnoga evaluation laboratory and compare pass rate against cost before choosing.
- Why is my output token count higher than expected?
- Thinking is enabled by default on this generation, and thinking tokens are billed as output. odnoga shows the real token split per request so the increase is visible rather than a surprise on the invoice.
All models · Pricing · Docs · Compare models in the evaluation lab
Sources: Anthropic — Claude Opus 5 overview, Anthropic — what is new in Claude Opus 5