Claude Sonnet 4.5 pricing.

Prices checked

Claude Sonnet 4.5 costs $3 in and $15 out per million tokens from Anthropic; the cheapest price is $5.15 at Poe, checked .

Anthropic's own price

Anthropic

$6/M tokens

$3 in · $15 out

581% above the typical price of the 392 text models we track.

Cheapest price

Poe

$5.15/M tokens

$2.58 in · $12.88 out

14% less than Anthropic, and the provider says why.

Web search from this model costs $10 per 1,000 searches, plus its tokens; dedicated search APIs from $1

tool callingreads imagesreasoningstructured output200k contextreleased 29 Sept 2025

Every provider selling Claude Sonnet 4.5

  1. 1

    Poe

    Credits

    $2.58 in · $12.88 out · 983k context here

    $5.15/M tokens
  2. 2

    Jiekou

    Resells

    $2.85 in · $14.25 out

    $5.70/M tokens
  3. 3

    Cortecs

    Same as maker

    $3.00 in · $14.99 out

    $6.00/M tokens
  4. 4

    OpenRouter

    Same as maker

    $3 in · $15 out · 1M context here

    $6/M tokens
  5. 5

    LLM Gateway

    Same as maker

    $3 in · $15 out

    $6/M tokens
  6. 6

    Vercel AI Gateway

    Same as maker

    $3 in · $15 out · 1M context here

    $6/M tokens
  7. 7

    OpenCode

    Same as maker

    $3 in · $15 out · 1M context here

    $6/M tokens
  8. 8

    OrcaRouter

    Same as maker

    $3 in · $15 out · 1M context here

    $6/M tokens
  9. 9

    Merge Gateway

    Same as maker

    $3 in · $15 out

    $6/M tokens
  10. 10

    $3 in · $15 out

    $6/M tokens
  11. 11

    Snowflake Cortex

    Same as maker

    $3 in · $15 out

    $6/M tokens
  12. 12

    Google Vertex AI

    Same as maker

    $3 in · $15 out

    $6/M tokens
  13. 13

    Kilo Gateway

    Same as maker

    $3 in · $15 out · 1M context here

    $6/M tokens
  14. 14

    Azure AI Foundry

    Same as maker

    $3 in · $15 out

    $6/M tokens
  15. 15

    Perplexity Agent

    Same as maker

    $3 in · $15 out

    $6/M tokens
  16. 16

    Helicone

    Same as maker

    $3 in · $15 out

    $6/M tokens
  17. 17

    Abacus

    Same as maker

    $3 in · $15 out

    $6/M tokens
  18. 18

    Anthropic

    Maker's price

    $3 in · $15 out

    $6/M tokens
  19. 19

    Tempr Gateway

    Same as maker

    $3 in · $15 out

    $6/M tokens
  20. 20

    Impossibl

    Same as maker

    $3 in · $15 out

    $6/M tokens
  21. 21

    Ofox

    Same as maker

    $3 in · $15 out

    $6/M tokens
  22. 22

    NEAR AI Cloud

    Same as maker

    $3 in · $15 out

    $6/M tokens
  23. 23

    Requesty

    Same as maker

    $3 in · $15 out · 1M context here

    $6/M tokens
  24. 24

    ZenMux

    Same as maker

    $3 in · $15 out · 1M context here

    $6/M tokens
  25. 25

    NanoGPT

    Same as maker

    $3 in · $15 out

    $6/M tokens

Not ranked

Below the maker's own price with nothing published to account for it, from a catalog we could not read reliably, or a free tier, which is a limit rather than a price. Why

  1. QiHang

    Not ranked

    $0.43 in · $2.14 out

    $0.86/M tokens
  2. Meganova

    Not ranked

    $2.40 in · $12 out · 1M context here

    $4.80/M tokens

Something else wrong?

What it costs for your work
pages a month
$2.70

A month at Anthropic's own price

$2.32

A month at Poe, the cheapest

$0.38

What you save

A page is about 600 tokens in and 60 out, a support reply 200 in and 300 out, an email 150 in and 250 out. Nothing else is counted: no cache discount, no batch rate, no free tier.

Evidence

Each is a measurement of the stated version and configuration. Two benchmarks are not the same test, and a score is not the chance your own work succeeds.

SWE-bench Verified 500

71.4 / 100

SWE-bench · model claude-sonnet-4-5-20250929 · mini-SWE-agent 2.0.0; attempts 1; reasoning high

Evaluated 17 Feb 2026, read 31 Oct 2026

Read the source

BFCL V4 multi-turn V4 / bfcl-eval 2025.12.17

61.37 / 100

Berkeley BFCL · model Claude-Sonnet-4-5-20250929 · FC; multi-turn

Published 17 Oct 2026, read 31 Oct 2026

Read the source

BFCL V4 web-search V4 / bfcl-eval 2025.12.17

81 / 100

Berkeley BFCL · model Claude-Sonnet-4-5-20250929 · FC; web-search

Published 17 Oct 2026, read 31 Oct 2026

Read the source

BFCL V4 memory V4 / bfcl-eval 2025.12.17

64.95 / 100

Berkeley BFCL · model Claude-Sonnet-4-5-20250929 · FC; memory

Published 17 Oct 2026, read 31 Oct 2026

Read the source
How we count

Prices are per million tokens. The big number on each row is blended 3 in : 1 out, the way most work uses a model. A price below Anthropic's own is listed as cheapest only when the provider says why.

Credits
Priced in credits or points that convert to dollars, so the rate moves when the conversion does.
Resells
The provider sells another company's endpoint, at a margin of its own or at the maker's price.
Same as maker
Sold by someone else at the price the maker charges.
Maker's price
What the lab that trained the model charges on its own API. Every other row is measured against it.
Costs more
Sold by someone else for more than the maker charges.
Not ranked
Below the maker's price on a closed model with nothing to account for it, or from a catalog we could not read reliably. Listed, and never a pick.

Similar models

Models from other labs that do what Claude Sonnet 4.5 does, cheapest first. Each percentage is against Claude Sonnet 4.5's cheapest price, $5.15 per M tokens.

Frequently asked

Who sells Claude Sonnet 4.5 cheapest?
Poe, at $2.58 in and $12.88 out per million tokens, checked 31 Oct 2026. Priced in credits or points that convert to dollars, so the rate moves when the conversion does.
Is Claude Sonnet 4.5 open weights?
No. Anthropic keeps the weights, so every price on this page is Anthropic's own or a resale of it. A price below Anthropic's here needs a published reason, or it is listed and not ranked.
How much context does Claude Sonnet 4.5 hold?
200,000 tokens in one call, which is about 200k, and it can call your own functions.
How does Claude Sonnet 4.5 compare with the rest of the market?
At $6 per million tokens, Anthropic's own price is 581% above the typical $0.88 across the 392 text models we track.

More from Anthropic