Last verified
FRONTIER LABFOUNDED 2021ANTHROPIC PBCCONSTITUTIONAL AIMCP CREATOR

Claude API Pricing

The AI safety-focused lab behind Claude. Founded by ex-OpenAI researchers in 2021, it ships frontier-grade reasoning models with zero-retention defaults and the most rigorous published RSP (Responsible Scaling Policy) in the field.

Production models
9
9 GA
Founded
2021
5 years operating
Max context
1M
Claude Sonnet 5
Cheapest tier
$1/M
Claude Haiku 4.5 input
Data retention
0days
Pro+ default
Headquarters
San Francisco
PBC · founded 2021
§ 01 / LINEUP

The full roster.

Side-by-side →
BALANCED-AGENTIC

Claude Sonnet 5

Official Anthropic Claude Sonnet page lists Claude Sonnet 5 introductory API pricing at $2/M input and $10/M output through 2026-08-31, with standard $3/M input and $15/M output starting 2026-09-01.

Input
$2/M
Output
$10/M
1M ctx · vision
MYTHOS

Claude Fable 5 LAUNCHED JUN 2026

Claude Fable 5 is listed by Anthropic at $10/M input, $12.50/M 5m cache writes, $20/M 1h cache writes, $1/M cache hits, and $50/M output.

Input
$10/M
Output
$50/M
1M ctx · vision
FRONTIER · REASONING

Claude Opus 4.8 LAUNCHED MAY 2026

Newest generally available Opus release.

Input
$5/M
Output
$25/M
1M ctx · vision
FRONTIER · REASONING

Claude Opus 4.7

Newest Opus.

Input
$5/M
Output
$25/M
1M ctx · vision
FRONTIER · REASONING

Claude Opus 4.6

.

Input
$5/M
Output
$25/M
1M ctx · vision
BALANCED · MID-TIER

Claude Sonnet 4.6

.

Input
$3/M
Output
$15/M
1M ctx · vision
FRONTIER · REASONING

Claude Opus 4.5

Repriced down from $15 input / $75 output launch list to match 4.6/4.7 tier.

Input
$5/M
Output
$25/M
200K ctx · vision
BALANCED · MID-TIER

Claude Sonnet 4.5

.

Input
$3/M
Output
$15/M
200K ctx · vision
FAST · LIGHTWEIGHT

Claude Haiku 4.5

.

Input
$1/M
Output
$5/M
200K ctx · text-only
RESTRICTED · NOT PUBLIC

Claude Mythos 5 LAUNCHED JUN 2026

⚠ Glasswing partners first; broader trusted access planned. No self-serve API alias published in launch post.. Claude Mythos 5 is listed by Anthropic at $10/M input, $12.50/M 5m cache writes, $20/M…

Input
$10/M
Output
$50/M
1M ctx · vision
DEPRECATED · 2026-07-11

Claude Opus 4.1

Legacy Opus pricing tier at $15/M input, $1.50/M cache-hit, $75/M output.

Input
$15/M
Output
$75/M
200K ctx · vision
DEPRECATED · 2026-04-14

Claude Opus 4 LAUNCHED MAY 2025

Deprecated on April 14, 2026 with retirement scheduled for June 15, 2026.

Input
$15/M
Output
$75/M
200K ctx · vision
DEPRECATED · 2026-04-14

Claude Sonnet 4 LAUNCHED MAY 2025

API model name claude-sonnet-4-20250514.

Input
$3/M
Output
$15/M
200K ctx · vision
RETIRED · FEB 2026

Claude Sonnet 3.7

Archive-only model.

Input
$3/M
Output
$15/M
200K ctx · vision
RETIRED · OCT 2025

Claude Sonnet 3.5 (October 2024) LAUNCHED OCT 2024

API model name claude-3-5-sonnet-20241022.

Input
$3/M
Output
$15/M
200K ctx · vision
RETIRED · FEB 2026

Claude Haiku 3.5

Retired from first-party Claude API on Feb 19, 2026 (still available on Amazon Bedrock and Vertex AI per their own deprecation schedules).

Input
$0.80/M
Output
$4/M
200K ctx · text-only
RETIRED · OCT 2025

Claude Sonnet 3.5 (June 2024) LAUNCHED JUN 2024

API model name claude-3-5-sonnet-20240620.

Input
$3/M
Output
$15/M
200K ctx · vision
RETIRED · JAN 2026

Claude Opus 3 LAUNCHED MAR 2024

API model name claude-3-opus-20240229.

Input
$15/M
Output
$75/M
200K ctx · vision
RETIRED · JUL 2025

Claude Sonnet 3 LAUNCHED MAR 2024

API model name claude-3-sonnet-20240229.

Input
$3/M
Output
$15/M
200K ctx · vision
RETIRED · APR 2026

Claude Haiku 3 LAUNCHED MAR 2024

API model name claude-3-haiku-20240307.

Input
$0.25/M
Output
$1/M
200K ctx · vision
RESTRICTED · NOT PUBLIC

Claude Mythos Preview

⚠ Glasswing partners only — no public API. Restricted-access Mythos Preview archive.

Input
$25/M
Output
$125/M
documented elsewhere ctx · text-only
§ 02 / SHELF

All side-by-side.

Methodology →
Model Input /M Output /M Cached Context Max output Vision Tools Tier
Claude Sonnet 5 $2.00 $10 $0.2−90% 1M Active
Claude Fable 5 FLAGSHIP $10 $50 $1.00−90% 1M Active
Claude Opus 4.8 $5.00 $25 $0.5−90% 1M Frontier
Claude Opus 4.7 $5.00 $25 $0.5−90% 1M Frontier
Claude Opus 4.6 $5.00 $25 $0.5−90% 1M Frontier
Claude Sonnet 4.6 $3.00 $15 $0.3−90% 1M Mid
Claude Opus 4.5 $5.00 $25 $0.5−90% 200K Frontier
Claude Sonnet 4.5 $3.00 $15 $0.3−90% 200K Mid
Claude Haiku 4.5 $1.00 $5.00 $0.1−90% 200K Light
Claude Mythos 5 RESTRICTED $10 $50 $1.00−90% 1M Glasswing partners first; broader trusted access planned. No self-serve API alias published in launch post. Restricted
Claude Opus 4.1 DEPRECATED $15 $75 $1.50−90% 200K Deprecated
Claude Opus 4 DEPRECATED $15 $75 $1.50−90% 200K Deprecated
Claude Sonnet 4 DEPRECATED $3.00 $15 $0.3−90% 200K Deprecated
Claude Sonnet 3.7 RETIRED Retired Feb 19, 2026 → claude-sonnet-4-6 Retired
Claude Sonnet 3.5 (October 2024) RETIRED Retired Oct 28, 2025 → claude-sonnet-4-6 Retired
Claude Haiku 3.5 RETIRED Retired Feb 19, 2026 → claude-haiku-4-5 Retired
Claude Sonnet 3.5 (June 2024) RETIRED Retired Oct 28, 2025 → claude-sonnet-4-6 Retired
Claude Opus 3 RETIRED Retired Jan 5, 2026 → claude-opus-4-7 Retired
Claude Sonnet 3 RETIRED Retired Jul 21, 2025 → claude-sonnet-4-6 Retired
Claude Haiku 3 RETIRED Retired Apr 20, 2026 → claude-haiku-4-5 Retired
Claude Mythos Preview RESTRICTED $25 $125 documented elsewhere Glasswing partners only — no public API Restricted
§ 03 / PRICE CURVE

Pricing across the lineup.

How Anthropic priced 15 models · MAR 24 → JUN 26.

oldest → newest →
$100$75$50$25 Claude Sonnet 3 MAR 24 Claude Haiku 3 MAR 24 Claude Sonnet 3… OCT 24 Claude Sonnet 4 MAY 25 Claude Opus 4.5 MAR 26 Claude Sonnet 4… APR 26 Claude Fable 5 JUN 26 Claude Sonnet 5 JUN 26
Input · newest $2/M
Output · newest $10/M
Each point is a model at its listed $/M price.

The last 12 months at Anthropic.

Major releases, pricing moves, and capability updates. Sourced from official Anthropic announcements + our own verification.

JUN 30 - 2026
Claude Sonnet 5 released - introductory API price $2/M input and $10/M output through Aug 31; standard $3/$15 starts Sep 1
RELEASE
JUN 09 · 2026
Claude Fable 5 and Mythos 5 launched · Mythos-class general release plus restricted trusted-access model at $10/$50
RELEASE
MAY 28 · 2026
Claude Opus 4.8 released — full 1M context at standard $5/$25 pricing, plus fast mode at $10/$50
RELEASE
MAY 17 · 2026
Opus 4.5 verified at $5/$25 per M (was $15/$75 at launch) — aligned with Opus 4.6/4.7; Anthropic does not publish a cut date
PRICING
APR 22 · 2026
Claude Sonnet 4.6 released — replaces 3.7 as default in Pro tier
RELEASE
APR 02 · 2026
Claude Haiku 4.5 launched at $1/M input — cheapest in lineup
RELEASE
MAR 18 · 2026
Prompt caching shipped — 90% off cached input tokens, 5min TTL
FEATURE
MAR 04 · 2026
Claude Opus 4.5 released — new flagship, leads SWE-bench at 81.3%
RELEASE
FEB 14 · 2026
Files API shipped — direct PDF + image upload to API without base64
FEATURE
JAN 08 · 2026
Voyage AI integrated — embedding model now available via Anthropic API
FEATURE
DEC 11 · 2025
MCP 1.0 spec finalized — Model Context Protocol becomes industry-adopted
FEATURE
SEP 22 · 2025
Voyage AI acquired — Anthropic gets best-in-class embedding capability
CORPORATE
MAR · 2025
Claude Code launched — terminal-native coding agent, immediate hit
RELEASE
§ 04 / ACCESS

Where to get it.

Methodology →
§ 04 / BEST FOR

Which Anthropic for what.

More scenarios →
If you need the most capable generally available Claude?
Claude Fable 5
Profile →
If you need frontier coding agents + complex reasoning…
Opus 4.8
Profile →
If you want the new balanced daily driver while intro pricing is active...
Sonnet 5
Profile ->
If you're doing high-volume classification or simple chat…
Haiku 4.5
Profile →
If you live in a terminal and want a coding agent…
Claude Code
Profile →
If you're building RAG / semantic search
Voyage 3 Large
Profile →
If you need EU data residency for compliance…
Opus via Bedrock EU
AWS docs →
§ 06 / BACKGROUND

The company behind it.

www.anthropic.com →

Anthropic was founded in 2021 by Dario and Daniela Amodei alongside several former OpenAI researchers — including Tom Brown (lead author of the GPT-3 paper) and the team behind Constitutional AI research. The pitch: build frontier-grade AI with safety and interpretability as first-class concerns, not afterthoughts.

The company is structured as a public-benefit corporation (PBC), which legally binds its directors to weigh public good alongside shareholder return. This is the strongest legal commitment to safety among frontier labs — a meaningful difference vs OpenAI's capped-profit structure or Google DeepMind's research-arm-of-Alphabet model.

Anthropic's flagship technical contributions: Constitutional AI (training models with explicit principles rather than only RLHF), the Responsible Scaling Policy (public commitments about which capabilities trigger which safety reviews), and Model Context Protocol (MCP — open spec for tool-using agents, now widely adopted across the industry).

Funding has come primarily from Google ($2B) and Amazon ($4B) — both also resell Claude through their cloud platforms. The latest valuation rounds put the company at $60B+ as of 2025. Revenue trajectory is in the multi-billion-dollar range, primarily from API customers and Claude consumer subscriptions.

Compared to OpenAI: smaller, narrower model lineup, more conservative on multimodal, but stronger on coding and reasoning benchmarks since 2024. Compared to Google: more focused product set, less ecosystem reach, but better SWE-bench scores. Direct head-to-head with GPT-5.

§ 07 / COMPETITORS

Other frontier labs.

All providers →

Frequently asked.

Practical questions about Anthropic and the Claude lineup.

Q · 01 What is the newest Claude model? +
The newest public Claude model on AI//COST is Claude Sonnet 5, released June 30, 2026. Anthropic lists an introductory API price of $2/M input and $10/M output through August 31, 2026; standard Sonnet $3/$15 pricing starts September 1, 2026.
Q · 02 Which Claude model should I start with? +
For most new API use cases: Claude Sonnet 5. Its introductory API rate is $2/M input and $10/M output through August 31, 2026, then it moves to standard Sonnet $3/$15 pricing. Use Opus for harder reasoning, Haiku for cheaper high-volume classification, and Fable/Mythos only where access and safeguards fit.
Q · 03 Is Anthropic safer than OpenAI? +
By stated policies: yes — Anthropic publishes a more detailed Responsible Scaling Policy and is structured as a PBC (legally bound to public good). By outcomes: roughly comparable; both labs ship frontier models, both have leak/red-team incidents. The structural difference matters most for compliance + procurement teams looking for documented safety commitments.
Q · 04 Does Anthropic offer EU data residency? +
Direct API: no — runs on US infrastructure with EU SCCs. For EU-only data: use Claude on AWS Bedrock in eu-central-1 (Frankfurt) or eu-west-3 (Paris). Same model, same pricing, but data never leaves the EU region. Enterprise contracts include custom DPAs.
Q · 05 What's the data retention policy? +
API requests are not used for training by default. Pro/Max/Enterprise tiers default to zero retention — your prompts are not stored. Free tier retains for 30 days for abuse monitoring only. This is the strongest privacy posture of any frontier US lab.
Q · 06 Will Claude 5 / Opus 5 launch this year? +
Anthropic hasn't announced a Claude 5. Their cadence has been roughly: major version every 12-18 months, minor versions (4.5 → 4.6 → 4.7) every 4-6 weeks. Opus 4.6 and 4.7 are already live alongside 4.5; Claude 5 unlikely before 2027 based on historical timing.
Q · 07 Can I get a discount? +
Yes via committed-spend Enterprise contracts — typically 10-30% off list price at $50K+/year commitments. For smaller users: prompt caching ($0.50/M input vs $5/M base on Opus 4.8) is the biggest realistic discount, no contract needed. AWS Bedrock + GCP Vertex offer their own committed-use discounts on top.
Q · 08 How does Anthropic compare to other frontier labs? +
Strongest at coding, reasoning, long-form writing, tool use. Weaker at multimodal (no native image gen, voice still English-only). After the 4.5-generation repricing, Opus 4.5/4.6/4.7/4.8 sit at $5/M input · $25/M output — competitive with GPT-5 and below Gemini 2.5 Pro on output cost. See comparisons.