BALANCED REASONING1M CONTEXTTEXT + VISIONPROMPT CACHINGBATCH -50%
Claude Sonnet 4.6 API Pricing
Anthropic's mainstream hybrid reasoning model for agents, coding, and production workflows — $3/M input, $15/M output, with cache reads at $0.30/M. It keeps the Sonnet price band while adding a 1M context window in API beta. Pulled directly from platform.claude.com daily.
Input · per 1M tokens
$3.00/M
Stable 25 days ↔ flat
Output · per 1M tokens
$15.00/M
Stable 25 days ↔ flat
Cached input · 90% off
$0.30/M
Cache 5min or 1h −90%
Effective · agentic blend
$1.92/M
92/8 split · 82% cache
§ 01 / TERMINAL
Run the numbers.
Live calculator pre-loaded with current Claude Sonnet 4.6 rates. Tweak spend, output mix, or cache hit rate — share the URL to share the calculation.
$ /mo
Workload split
Prompt cache hit rate
Tokens you can process
—
Words equivalent (English)
—
Effective rate
—
§ 02 / SCENARIOS
Real-world presets.
CUSTOMER SUPPORT
Support agent ticket triage
$0.011/ticket
CODING AGENT
Repo-wide bug fix
$0.169/task
LONG DOC ANALYSIS
Reading 100-page contracts
$0.645/doc
CONTENT OPS
Editorial brief draft
$0.048/brief
§ 03 / TOKENIZER
Paste text. See tokens. See cost.
Estimate · anthropic-bpe-estimate · ≈3.5 chars/token Auto-counts as you type
This is a chars-per-token approximation, not a real tokenizer. Actual tokens vary by language, code density, and tool-call overhead — counts are typically ±10–20% off for English prose, more for code or non-Latin scripts. For exact billing, use the vendor's official tokenizer.
Characters 521
Words 79
Tokens (estimated) 149 tokens
Cost as input · uncached $0.000447 USD
Cost as output · uncached $0.002235 USD
Cost as cached input $0.000045 USD
| Model | Input /M | Output /M | Effective blended | Context | Best for |
|---|---|---|---|---|---|
| Claude Sonnet 4.6 Current | $3.00 cache $0.30 | $15.00 | $1.92 agentic 92/8 | 1M | Production agents · coding |
| Claude Opus 4.7 | $5.00 cache $0.50 | $25.00 | $3.21 pricier | 1M | Frontier reasoning · hard code |
| Claude Sonnet 4.5 | $3.00 cache $0.30 | $15.00 | $1.92 same list price | 200K | Stable Sonnet migrations |
| Claude Haiku 4.5 | $1.00 cache $0.10 | $5.00 | $0.64 cheaper | 200K | Support · classification |
| GPT-5.4 | $2.50 cache $0.25 | $15.00 | $1.80 near parity | 270K | Affordable coding · tool use |
| Gemini 2.5 Pro | $1.25 cache $0.13 | $10.00 | $1.10 cheaper | 2M | Long-context document analysis |
| DeepSeek V4 Flash | $0.14 cache $0.00 | $0.28 | $0.05 deep budget | 1M | Bulk RAG · low-cost generation |
| DeepSeek V4 Pro | $0.43 cache $0.00 | $0.87 | $0.14 75% off through May 31 | 1M | Math · reasoning · code analysis |
Frequently asked.
Practical pricing questions, with the live vendor numbers separated from workload assumptions.
Q · 01 Is Claude Sonnet 4.6 the most expensive mid-tier model? +
No. At
$3/M input and $15/M output, Sonnet 4.6 is cheaper than Claude Opus but pricier than GPT-5, Gemini 2.5 Pro, and DeepSeek-class budget models on raw token rates. With an 82% cache hit rate and a 92/8 input-output workload, its effective blended cost is $1.92/M.Q · 02 How does prompt caching change the price? +
Anthropic lists Sonnet 4.6 cache hits at
$0.30/M, which is 10% of the $3/M base input price. Cache writes are charged separately: $3.75/M for 5-minute writes and $6/M for 1-hour writes. The page calculator uses cache-hit pricing for repeated prompt sections, not for first-time cache writes.Q · 03 When was the last price change? +
We did not find a posted price change after the lineup launch date used on AI//COST. Anthropic's current pricing table lists Claude Sonnet 4.6 at
$3/M input, $0.30/M cache hits, and $15/M output, the same headline price tracked from April 22, 2026.Q · 04 Does Sonnet 4.6 have a long-context surcharge? +
Anthropic says Claude Sonnet 4.6 includes the full
1M token context window at standard pricing. That means a 900k-token request uses the same per-token list rate as a shorter request, before any prompt caching, batch, or data-residency modifiers.Q · 05 Does pricing differ by region? +
The default first-party Claude API route uses standard global pricing. For Claude Sonnet 4.6 and later models, setting
inference_geo: "us" applies a 1.1x multiplier to input, output, cache write, and cache read pricing. Bedrock and Vertex AI have their own regional pricing policies.Q · 06 Is there a Batch API discount? +
Yes. Anthropic's Batch API gives a 50% discount on both input and output tokens. For Sonnet 4.6, the batch table lists
$1.50/M input and $7.50/M output, compared with the standard $3/M and $15/M rates.Q · 07 Are volume discounts published? +
No fixed volume-discount ladder is published. Anthropic says volume discounts may be available for high-volume users and are negotiated case by case. Standard API users should assume the public rates shown here unless they have a separate enterprise agreement.
Q · 08 How accurate is the tokenizer estimate? +
The live counter uses a
4.875 English characters-per-token estimate for Claude-class BPE tokenization. It is useful for planning but billing is based on Anthropic's server-side token count. Non-English text, code, and tables can deviate materially from the English estimate.