Input
- Official Anthropic
- $3
- Here (−50%)
- $1.5
claude-sonnet-4-6
Claude Sonnet 4.6 API — price per token
Claude Sonnet 4.6 is the previous-generation balanced model — a proven workhorse for coding and production pipelines. Sonnet 5 is cheaper at $2/$10.
Pricing per 1M tokens
| Rate | Official Anthropic | Here (−50%) |
|---|---|---|
| Input | $3 | $1.5 |
| Output | $15 | $7.5 |
| Cache read | $0.3 | $0.15 |
| Cache write (5m) | $3.75 | $1.875 |
Output
- Official Anthropic
- $15
- Here (−50%)
- $7.5
Cache read
- Official Anthropic
- $0.3
- Here (−50%)
- $0.15
Cache write (5m)
- Official Anthropic
- $3.75
- Here (−50%)
- $1.875
Every request is metered at the official rate first, then the flat 50% B2C discount is subtracted before it touches your prepaid balance. Context window: 200K tokens. Max output: 128K tokens.
Best for
- Pipelines tuned and evaluated against Sonnet 4.6.
- Balanced coding and content workloads.
- Teams migrating gradually to Sonnet 5.
Good to know
- Sonnet 5 is cheaper at $2/$10 — new projects should usually start on claude-sonnet-5.
- Supports adaptive thinking; effort defaults to high.
- Context is 200K tokens on this endpoint. Anthropic returns a long-context rate_limit_error above roughly 200K on this model, so a longer document belongs on claude-sonnet-5 or claude-opus-4-6 — both answer the same document here (measured 2026-08-29).
How to use Claude Sonnet 4.6
Create a free account, generate one key, and point any Anthropic-compatible tool at https://router.apitoken.sale with model ID claude-sonnet-4-6. Eligible new accounts include $5 of platform bonus credit — enough to test the model before topping up. Existing integrations on the legacy host https://api.apitoken.sale keep working.
Frequently asked questions
How much does the Claude Sonnet 4.6 API cost?
Officially $3 per 1M input tokens and $15 per 1M output tokens. With the flat 50% apiToken.sale discount that is $1.50/$7.50.
Sonnet 4.6 or Sonnet 5?
They do not share a list price: Sonnet 5 is $2/$10, Sonnet 4.6 is $3/$15. Sonnet 5 is also stronger on coding and agentic work — prefer it for new projects. Stay on 4.6 when your prompts and evals are pinned to it.
How large a document can I send to Sonnet 4.6?
About 200K input tokens. Above that Anthropic answers a long-context rate_limit_error on this model, and no retry helps. Send longer documents to claude-sonnet-5 or claude-opus-4-6. Note that the same text counts about 35% more tokens on Sonnet 5 than on 4.6, because the newer generation uses a different tokenizer.
Can I switch models without a new key?
Yes. One key and one prepaid balance cover every supported Claude and GPT model — switching is a model-ID and endpoint change.
Run Claude Sonnet 4.6 on the same Anthropic API at a flat 50% off — instant key, prepaid balance, card or crypto.