Claude API pricing in 2026: Anthropic's rates vs QuickSilver Pro
Per-token prices for Claude Opus 5.5, Fable 5.1, Sonnet 5.5 and Haiku 5.5 — Anthropic's list price next to what the same model costs through QuickSilver Pro, and how to switch.
Every current Claude model is billed per token, separately for input (what you send) and output (what the model writes back). All prices below are in US dollars per 1 million tokens. The QuickSilver Pro column is read from our live pricing table when this page is built, so it is always the price you are actually charged.
Current Claude models
| Model | Anthropic list (input / output) | QuickSilver Pro (input / output) | You save |
|---|---|---|---|
| Claude Opus 5.5 | $4.00 / $20.00 | $1.00 / $5.00 | 75% |
| Claude Fable 5.1 | $10.00 / $50.00 | $2.50 / $12.50 | 75% |
| Claude Sonnet 5.5 | $2.00 / $10.00 | $0.50 / $2.50 | 75% |
| Claude Haiku 5.5 | $0.10 / $0.50 | $0.025 / $0.125 | 75% |
Anthropic's list prices are from anthropic.com/pricing.
Haiku 5.5 has one extra rule: when a single request's prompt is longer than 200K tokens, that request is billed at $0.125 / $0.625 instead. The other three models have one rate across their whole 1M-token context window.
Which one to use
- Claude Sonnet 5.5 is the sensible default for most work: coding agents, chat, extraction, writing.
- Claude Opus 5.5 is for the hard cases — a large refactor, a bug Sonnet keeps missing, long agentic runs — at $1.00 / $5.00.
- Claude Haiku 5.5 is for volume: classification, routing, summarising, and the background calls that agent tools make on your behalf.
- Claude Fable 5.1 is Anthropic's Mythos-class model for the deepest reasoning. It costs $2.50 / $12.50, so reach for it when Opus is not enough, not by default.
Output tokens cost five times as much as input tokens on every Claude model, so the length of the answer usually matters more than the length of the prompt. Asking for a diff instead of a whole rewritten file is the single biggest saving in coding work.
Older Claude models
These stay available for applications pinned to them:
| Model | Anthropic list (input / output) | QuickSilver Pro (input / output) | You save |
|---|---|---|---|
| Claude Opus 5 | $5.00 / $25.00 | $1.25 / $6.25 | 75% |
| Claude Fable 5 | $10.00 / $50.00 | $2.50 / $12.50 | 75% |
| Claude Opus 4.8 | $5.00 / $25.00 | $1.25 / $6.25 | 75% |
| Claude Sonnet 5 | $2.00 / $10.00 | $0.50 / $2.50 | 75% |
| Claude Haiku 4.5 | $1.00 / $5.00 | $0.25 / $1.25 | 75% |
How to switch
You need an API key from the dashboard — new accounts get $0.05 of free credit, no card required, which works with Claude Haiku 5.5 and the other starter models; the larger Claude models need a top-up, from $5. Then change two values in your client: the base URL and the key.
With the Anthropic SDK, keep your code and point it at QuickSilver Pro:
import anthropic
client = anthropic.Anthropic(
base_url="https://api.quicksilverpro.io",
api_key="YOUR_QSP_KEY",
)
message = client.messages.create(
model="claude-sonnet-5-5",
max_tokens=1024,
messages=[{"role": "user", "content": "Explain this stack trace: ..."}],
)
print(message.content[0].text)
Or with the OpenAI SDK, which works for every model in the catalog, Claude included:
from openai import OpenAI
client = OpenAI(
base_url="https://api.quicksilverpro.io/v1",
api_key="YOUR_QSP_KEY",
)
response = client.chat.completions.create(
model="claude-sonnet-5-5",
messages=[{"role": "user", "content": "Explain this stack trace: ..."}],
)
print(response.choices[0].message.content)
Model IDs use hyphens: claude-opus-5-5, claude-sonnet-5-5, claude-haiku-5-5, claude-fable-5-1. Billing is prepaid, so there is no surprise invoice — usage is deducted from your balance per request, and every request shows up in the dashboard with its token counts and cost.
Using Claude Code? See how to run Claude Code on QuickSilver Pro, with a measured cost per task.