Claude Sonnet 5
Claude Sonnet 5 by Anthropic: $2 per 1M input and $10 per 1M output tokens, 1M context. Call it with one OpenAI-compatible API key, compare it with similar models, start free.
Claude Sonnet 5 is served by Anthropic and called through odnoga on /v1/chat/completions, in the same OpenAI-compatible shape every other model on that endpoint uses. It is a reasoning model, so it spends extra output tokens working through a problem before answering — budget for higher output cost on hard tasks. It accepts images alongside text. It supports tool and function calling. It can be forced to return structured JSON. Its 1M-token context window sets how much input you can send in one request. Cached input is billed at $0.2 per 1M tokens, so repeated prefixes cost less.
Specification and price
| Vendor | Anthropic |
|---|---|
| Model ID | claude-sonnet-5 |
| Context window | 1M tokens |
| Max output | 128K tokens |
| Served on | /v1/chat/completions |
| Capabilities | Reasoning, Reads images, Tool calling, JSON schema, Streaming, Web search, Prompt caching, Batch |
| Per 1M tokens | Vendor cost | Your price on Free+7% |
|---|---|---|
| Input | $2 | $2.15 |
| Output | $10 | $10.75 |
| Cached input | $0.2 | $0.215 |
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-sonnet-5",
messages: [{ role: "user", content: "Hello" }],
}),
});
One endpoint, one key, one bill — /v1/chat/completions takes the same shape whichever vendor is behind it. Swap the model id and odnoga handles keys, routing, limits and cost accounting.
What Claude Sonnet 5 can do
Everything below is what the model itself supports. Ask for something it does not do and odnoga tells you before the request reaches the provider, instead of returning a worse answer.
Works through a problem step by step before answering, which costs more tokens and buys accuracy on hard tasks.
Send a screenshot, a photo or a scan in the same message as your text.
The model can call functions you define and use what they return.
Answers with valid JSON instead of prose you would have to parse.
Answers in the exact shape you specify, so the response fits your types without a fallback path.
Tokens arrive as they are produced, so your interface can answer immediately.
Can look something up rather than answer from training data alone. Billed per search.
A prompt you send repeatedly is charged at the cached rate, which is a fraction of the input price.
Work that can wait goes through at a lower price.
Beyond text
Send a document and ask about it, without extracting the text yourself first.
Finer control
- Structured outputs · The provider guarantees the schema rather than merely aiming for it.
- Parallel tools · Can call several of your functions in one turn instead of one after another.
How to use Claude Sonnet 5
Reviewed 2026-09-14
Best for
- The default production model for agentic products: Anthropic describes it as the best combination of speed and intelligence, built to plan, use browsers and terminals and run autonomously.
- Customer-facing assistants that also call tools.
- Long documents and large codebases — 1M-token context, 128K max output.
- Teams upgrading from Sonnet 4.6 who want the same integration shape.
Not the right pick when
- Simple bulk classification — Haiku-class models are cheaper for that.
- Code that sets temperature, top_p or top_k: on Sonnet 5 non-default sampling parameters return an error.
- The very hardest long-horizon reasoning, where Opus 5 or a Fable-class model may justify its price.
Practical tips through odnoga
- 01Remove temperature and top_p from your request before switching: Anthropic returns a 400 for non-default sampling values on this model, and manual extended thinking is no longer accepted.
- 02Adaptive thinking is on by default — treat thinking tokens as part of the output budget.
- 03Front-load the stable system context so prompt caching applies to the repeated part.
- 04Keep the previous model in the catalog and A/B the two versions in odnoga rather than cutting over blind; the ledger shows quality-neutral cost differences per tenant.
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) | $21.51 |
|---|---|
| Output (2M tokens) | $21.51 |
| Your cost per month on Free | $43.01 |
Vendor list cost $40.00 + odnoga margin $3.01 (+7%). Cached input or answer reuse lowers it; odnoga records both numbers per request.
If Claude Sonnet 5 is not the one
Models that take the same kind of input, are priced the same way, and do the closest thing to what this one does. Not a quality ranking — odnoga does not benchmark models.
| Model | Context | Input / 1M | Output / 1M | Capabilities |
|---|---|---|---|---|
| Claude Sonnet 5 | 1M | $2 | $10 | Reasoning, Reads images, Tool calling, JSON schema, Streaming, Web search, Prompt caching, Batch |
| Claude Sonnet 5.5 | 1M | $2 | $10 | Reasoning, Reads images, Tool calling, JSON schema, Streaming, Web search, Prompt caching, Batch |
| gpt-6-sol | 1.1M | $2 | $10 | Reasoning, Reads images, Tool calling, JSON schema, Streaming, Prompt caching, Batch |
| Gemini 3.1 Pro Preview | 1.0M | $2 | $12 | Reasoning, Reads images, Tool calling, JSON schema, Streaming, Web search, Prompt caching, Batch |
Questions
- Is Claude Sonnet 5 a drop-in upgrade from Sonnet 4.6?
- Anthropic calls it a drop-in upgrade with three behaviour changes: adaptive thinking on by default, manual extended thinking now returns an error, and non-default temperature, top_p or top_k return an error.
- Sonnet 5 or Opus 5?
- Start with Sonnet 5 for most production traffic and escalate only the prompts your evaluation shows need Opus 5. In odnoga that is a prompt-level decision, not an application rewrite.
- Does the 1M context mean I should send everything?
- No — you pay for every input token. Send what the task needs, keep the reusable prefix first so caching applies, and let the per-request ledger show what the context is actually costing you.
All models · Pricing · Docs · Compare models in the evaluation lab
Sources: Anthropic — Claude Sonnet 5 overview, Anthropic — introducing Claude Sonnet 5