Last updated: October 20, 2026 · Prices checked against Anthropic's official pricing page and Oxyy's live Claude pricing page on this date.
Disclosure: This guide is published by the Oxyy.ai team. Official prices are Anthropic's; Oxyy is an independent gateway, not affiliated with Anthropic.
Quick answer: Claude Opus 5.5 costs $4 per 1M input tokens and $20 per 1M output tokens, and Claude Sonnet 5.5 costs $2 / $10 on Anthropic's API. Claude Opus 5 is $5 / $25 and Sonnet 5 is $2 / $10. Prompt-cache hits on Opus 5.5 and Sonnet 5.5 cost just 5% of the input price. Through Oxyy.ai, the same models currently cost 60% less: Opus 5.5 at $1.60 / $8.00 and Sonnet 5.5 at $0.80 / $4.00.
| Claude model | Anthropic official (input / output per 1M) | Oxyy.ai 60% off | Cache hits (official → Oxyy) | Context |
|---|---|---|---|---|
| Claude Opus 5.5 | $4.00 / $20.00 | $1.60 / $8.00 | $0.20 → $0.08 | 1M |
| Claude Sonnet 5.5 | $2.00 / $10.00 | $0.80 / $4.00 | $0.10 → $0.08 | 1M |
| Claude Opus 5 | $5.00 / $25.00 | $2.00 / $10.00 | $0.50 → $0.20 | 1M |
| Claude Sonnet 5 | $2.00 / $10.00 | $0.80 / $4.00 | $0.20 → $0.08 | 1M |
| Claude Sonnet 4.6 | $3.00 / $15.00 | $1.20 / $6.00 | $0.30 → $0.12 | 1M |
| Claude Haiku 5.5 (prompts ≤100K) | $0.10 / $0.50 | $0.04 / $0.20 | $0.01 → $0.004 | 1M |
| Claude Haiku 4.5 | $1.00 / $5.00 | $0.40 / $2.00 | $0.10 → $0.04 | 200K |
| Claude Fable 5.1 | $10.00 / $50.00 | $4.00 / $20.00 | $0.25 → $0.10 | 1M |
Official prices: Anthropic pricing page (platform.claude.com). Oxyy prices: oxyy.ai/anthropic, with a current 60% "Launch Promo". Checked October 20, 2026.
Key Takeaways
- Opus 5.5 is cheaper than Opus 5. Opus 5.5 lists at $4 / $20 per 1M tokens, versus $5 / $25 for Opus 5 and Opus 4.5–4.8.
- Sonnet 5's $2 / $10 price is now permanent. Anthropic cancelled the planned September 1, 2026 increase to $3 / $15.
- Caching is the biggest lever. Cache hits cost 10% of input on most models, and just 5% on Opus 5.5 and Sonnet 5.5.
- Batch API saves 50% on input and output, and it stacks with prompt caching.
- The newer tokenizer matters. Claude 4.7 and later produce about 30% more tokens for the same text, so compare cost per task, not just price per token.
- Oxyy charges 40% of list price on every Claude model, served by Anthropic upstream.
Claude API Pricing: Full Official Price List (October 2026)
| Model | Input | 5-min cache write | 1-hour cache write | Cache hits | Output |
|---|---|---|---|---|---|
| Claude Fable 5.1 | $10 | $12.50 | $20 | $0.25 | $50 |
| Claude Fable 5 | $10 | $12.50 | $20 | $1.00 | $50 |
| Claude Opus 5.5 | $4 | $5 | $8 | $0.20 | $20 |
| Claude Opus 5 | $5 | $6.25 | $10 | $0.50 | $25 |
| Claude Opus 4.5 – 4.8 | $5 | $6.25 | $10 | $0.50 | $25 |
| Claude Sonnet 5.5 | $2 | $2.50 | $4 | $0.10 | $10 |
| Claude Sonnet 5 | $2 | $2.50 | $4 | $0.20 | $10 |
| Claude Sonnet 4.6 / 4.5 | $3 | $3.75 | $6 | $0.30 | $15 |
| Claude Haiku 5.5 (≤100K prompt) | $0.10 | $0.125 | $0.20 | $0.01 | $0.50 |
| Claude Haiku 5.5 (>100K prompt) | $0.50 | $0.625 | $1 | $0.05 | $2.50 |
| Claude Haiku 4.5 | $1 | $1.25 | $2 | $0.10 | $5 |
All prices per 1M tokens (MTok), in USD. Claude Mythos 5.1 and Mythos 5 are priced like Fable but limited to verified organizations.
Claude Opus 5.5 API Pricing
Claude Opus 5.5 (released September 22, 2026) is Anthropic's top generally available Opus model for complex agentic coding and enterprise work. It has a 1M-token context window at standard pricing.
- Input: $4.00 per 1M tokens (Oxyy: $1.60)
- Output: $20.00 per 1M tokens (Oxyy: $8.00)
- Cache hits: $0.20 per 1M, which is 0.05× input (Oxyy: $0.08)
- 5-minute cache write: $5.00 · 1-hour cache write: $8.00
- Batch: 50% off input and output
- Fast mode: available on the Claude API at premium pricing
Claude Sonnet 5.5 API Pricing
Claude Sonnet 5.5 (released September 28, 2026) is the speed-and-intelligence middle tier, and the default choice for most production apps and coding agents.
- Input: $2.00 per 1M tokens (Oxyy: $0.80)
- Output: $10.00 per 1M tokens (Oxyy: $4.00)
- Cache hits: $0.10 per 1M, which is 0.05× input (Oxyy: $0.08)
- 5-minute cache write: $2.50 · 1-hour cache write: $4.00
- Context: 1M tokens at standard pricing
How Claude Prompt Caching Pricing Works
Prompt caching lets Claude reuse a repeated prompt prefix, such as a long system prompt, a document or a tool list, instead of processing it again.
| Cache operation | Price multiplier (vs base input) |
|---|---|
| 5-minute cache write | 1.25× |
| 1-hour cache write | 2× |
| Cache hit (most models) | 0.1× |
| Cache hit (Opus 5.5, Sonnet 5.5) | 0.05× |
| Cache hit (Fable 5.1, Mythos 5.1) | 0.025× |
Bonus: on most Claude models, cached input tokens don't count toward your input-tokens-per-minute rate limit. With an 80% cache hit rate and a 2M ITPM limit, Anthropic says you can effectively process 10M input tokens per minute.
Other Claude Pricing Rules That Change Your Bill
- Batch API: 50% off input and output for asynchronous jobs. It combines with prompt caching.
- Long context: Claude 4.6 and later (except Haiku 5.5) include the full 1M-token context at standard pricing. A 900K-token request is billed at the same per-token rate as a 9K-token one.
- Haiku 5.5 long prompts: requests over 100,000 prompt tokens pay 5× the short-prompt price.
- Data residency:
inference_geo: "us"applies a 1.1× multiplier on Claude 4.6 and later. - Web search tool: $10 per 1,000 searches plus tokens.
- Tokenizer: Claude 4.7 and later use a newer tokenizer that produces about 30% more tokens for the same text. Sonnet 4.6 and earlier use the old one.
Claude API Cost Examples
| Workload | Model | Anthropic list price | Oxyy |
|---|---|---|---|
| Single request (2,000 in + 500 out) | Opus 5.5 | $0.0180 | $0.0072 |
| Chatbot: 10M in + 2M out per month | Sonnet 5.5 | $40.00 | $16.00 |
| SaaS app: 200M in + 20M out per month | Sonnet 5.5 | $600.00 | $240.00 |
| SaaS app: 200M in + 20M out per month | Opus 5.5 | $1,200.00 | $480.00 |
| Startup: $5,000/month at list price | Any Claude model | $5,000.00 | $2,000.00 |
That last row is $36,000 saved per year. Caching, batch and Oxyy's discount stack, so your real bill can drop much further.
Claude API Rate Limits by Tier
| Tier | Monthly spend cap | Opus 5.5 / Sonnet 5.5 RPM | Input tokens / min | Output tokens / min |
|---|---|---|---|---|
| Start | $500 | 1,000 | 2,000,000 | 400,000 |
| Build | $1,000 | 5,000 | 5,000,000 | 1,000,000 |
| Scale | $200,000 | 10,000 | 10,000,000 | 2,000,000 |
| Custom | None | Arranged with Anthropic | ||
New organizations may start in an Evaluation tier with lower limits. Through Oxyy, your limits are set by your Oxyy tier instead, and they rise automatically as you top up.
Which Claude Model Should You Use?
| Use case | Best model | Price on Oxyy (per 1M) |
|---|---|---|
| Hardest reasoning and long-horizon agents | Opus 5.5 | $1.60 / $8.00 |
| Coding agents and production apps | Sonnet 5.5 | $0.80 / $4.00 |
| High-volume classification, routing, extraction | Haiku 5.5 | $0.04 / $0.20 |
| Maximum capability, cost no object | Fable 5.1 | $4.00 / $20.00 |
| Legacy prompts tuned for the old tokenizer | Sonnet 4.6 | $1.20 / $6.00 |
How to Pay 60% Less for Claude
Oxyy.ai is a unified AI API gateway that currently charges 40% of the standard price on every Claude rate. Claude requests are served by Anthropic upstream, and Oxyy supports both the Anthropic SDK and the OpenAI SDK.
Anthropic SDK (Python)
import anthropic, os
client = anthropic.Anthropic(
api_key=os.environ["OXYY_API_KEY"],
base_url="https://api.oxyy.ai", # no /v1: the SDK adds it
)
message = client.messages.create(
model="claude-sonnet-5-5",
max_tokens=1024,
messages=[{"role": "user", "content": "Hello, Claude"}],
)
print(message.content[0].text)OpenAI SDK
from openai import OpenAI
import os
client = OpenAI(base_url="https://api.oxyy.ai/v1", api_key=os.environ["OXYY_API_KEY"])
resp = client.chat.completions.create(model="claude-opus-5-5", messages=[{"role": "user", "content": "Hello"}])Claude Code: set ANTHROPIC_BASE_URL=https://api.oxyy.ai and put your Oxyy key in ANTHROPIC_AUTH_TOKEN.
What you get with Oxyy:
- 60% off input and output on every Claude model (current launch promo), and 60% off cache hits on most models. The exception is Sonnet 5.5 cache hits: $0.08 on Oxyy versus $0.10 at Anthropic.
- Per-block
cache_controlpreserved, extended thinking supported, and a freecount_tokensendpoint that doesn't use your rate limit - Live latency, throughput and 30-day success rate on every model page
- No subscription, no monthly minimum, credits never expire, and a free starting balance
- One key for Claude plus GPT, Gemini, Grok, GLM, DeepSeek and Kimi
Final Verdict
In 2026, Claude got cheaper at the top: Opus 5.5 at $4 / $20 undercuts Opus 5, and Sonnet 5.5 at $2 / $10 is now the standard mid-tier price. Use caching (5% hit price on the 5.5 models) and batch (50% off) wherever you can. Then cut the rest by 60% by routing through Oxyy, where Opus 5.5 costs $1.60 / $8.00 and Sonnet 5.5 costs $0.80 / $4.00.
Get Claude Opus 5.5 and Sonnet 5.5 at 60% off with a free Oxyy balance →
Related: All Claude prices on Oxyy · Claude Opus 5 · Claude Sonnet 5 · Prompt caching docs
Frequently Asked Questions
How much does the Claude Opus 5.5 API cost?
Claude Opus 5.5 costs $4 per 1M input tokens and $20 per 1M output tokens on Anthropic's API, with cache hits at $0.20 per 1M. On Oxyy it currently costs $1.60 / $8.00.
How much does the Claude Sonnet 5.5 API cost?
Claude Sonnet 5.5 costs $2 per 1M input tokens and $10 per 1M output tokens, with cache hits at $0.10 per 1M. On Oxyy it currently costs $0.80 / $4.00.
Is Claude Opus 5.5 cheaper than Opus 5?
Yes. Opus 5.5 lists at $4 / $20 per 1M tokens, while Opus 5 and Opus 4.5 through 4.8 list at $5 / $25.
Did Claude Sonnet 5's price go up to $3 / $15?
No. Anthropic announced $2 / $10 as introductory pricing through August 31, 2026, then made it the standard price. The planned increase to $3 / $15 will not happen.
How much does Claude prompt caching cost?
5-minute cache writes cost 1.25× the input price and 1-hour writes cost 2×. Cache hits cost 0.1× input on most models, 0.05× on Opus 5.5 and Sonnet 5.5, and 0.025× on Fable 5.1.
Does the Claude Batch API save money?
Yes. The Batch API gives a 50% discount on input and output tokens, and it can be combined with prompt caching.
Do long prompts cost more on Claude?
Not on Claude 4.6 and later, except Haiku 5.5. These models include the full 1M-token context at standard pricing. Haiku 5.5 charges higher rates for prompts over 100,000 tokens.
What is the cheapest way to use the Claude API?
Combine prompt caching and batch processing with a lower per-token price. Oxyy.ai currently charges 40% of Anthropic's list price on every Claude model.
Are Claude models on Oxyy the real Anthropic models?
Yes. Oxyy's Claude model pages list Anthropic as the upstream provider. Oxyy is an independent gateway and isn't affiliated with Anthropic.
Why does my Claude bill look higher on newer models?
Claude 4.7 and later use a newer tokenizer that produces about 30% more tokens for the same text, so the same prompt can cost more tokens even when the per-token price is lower.
What are Claude API rate limits on the Start tier?
On the Start tier, Opus 5.5 and Sonnet 5.5 allow 1,000 requests per minute, 2,000,000 input tokens per minute and 400,000 output tokens per minute, with a $500 monthly spend cap.