AI model · Anthropic

Claude Opus 5.2: Not Released. Meet Opus 5.5 Instead

Q: What is the latest Claude Opus model?

Claude Opus 5.5, with the API ID claude-opus-5-5, released September 22, 2026. It has a 1M-token context window, returns up to 128K tokens, and costs $4 per million input tokens and $20 per million output tokens. Anthropic marks it as the latest Opus and recommends it as the starting point for most coding and knowledge work.

Q: What model ID should I use instead of claude-opus-5-2?

Use claude-opus-5-5 on the Claude API, Google Cloud, Microsoft Foundry and Claude Platform on AWS, or anthropic.claude-opus-5-5 on Amazon Bedrock. A request for claude-opus-5-2 will fail because no such model exists. If you need more capability than Opus 5.5, the next step up is claude-fable-5-2.

Claude Opus 5.2 does not exist: as of November 14, 2026, Anthropic's newest Opus is Opus 5.5, released September 22, 2026 at $4 per million input tokens. Synexa does not serve Claude; use Synexa when your app also needs images, video, audio or 3D from one API key.

Current Opus: 5.5, $4 / $20 per MTok
The 5.2 models are Fable and Mythos
Synexa: image, video, audio, 3D APIs

Get Started

None

Claude Opus 5.2 releases from Anthropic

Sep 22, 2026

Release date of Opus 5.5, the latest Opus

$4 / $20

Opus 5.5 input / output price per million tokens

40%

Lower cost than Opus 5 on typical workloads, per Anthropic

Is there a Claude Opus 5.2?

No. Anthropic has not announced or shipped a model called Claude Opus 5.2. It is missing from Anthropic's models overview, from the model deprecations table that lists every Opus ID, and from the API release notes, and the documentation has no Opus 5.2 page. That is true as of November 14, 2026.

The Opus line went straight from Opus 5, released July 24, 2026, to Opus 5.5 on September 22, 2026. Opus 5.5 is labelled "Latest" in the docs, and Opus 5 is now a legacy model.

The "5.2" number does exist, just not on Opus. On November 9, 2026 Anthropic released Claude Fable 5.2 and Claude Mythos 5.2, its top models for the hardest technical work. So a search for Claude Opus 5.2 usually means one of two things: the current Opus, which is 5.5, or the new 5.2 models. This page covers both. For the 5.2 models in depth, read our Fable 5.2 and Mythos 5.2 guides.

Which Claude Opus is the latest?

Every Opus version Anthropic lists as active, newest first, as of November 2026.

Opus 5.5

API ID: claude-opus-5-5
Released: September 22, 2026
Status: Active, latest

Opus 5

API ID: claude-opus-5
Released: July 24, 2026
Status: Active, legacy

Opus 4.8

API ID: claude-opus-4-8
Released: not compared
Status: Active, legacy

Opus 4.7

API ID: claude-opus-4-7
Released: not compared
Status: Active, legacy

Opus 4.6

API ID: claude-opus-4-6
Released: not compared
Status: Active, legacy

Opus 4.5

API ID: claude-opus-4-5-20251101
Released: not compared
Status: Retirement not sooner than November 24, 2026

Claude Opus 5.2

API ID: none
Released: not released
Status: No such model

Model	API ID	Released	Status
Opus 5.5	`claude-opus-5-5`	September 22, 2026	Active, latest
Opus 5	`claude-opus-5`	July 24, 2026	Active, legacy
Opus 4.8	`claude-opus-4-8`	not compared	Active, legacy
Opus 4.7	`claude-opus-4-7`	not compared	Active, legacy
Opus 4.6	`claude-opus-4-6`	not compared	Active, legacy
Opus 4.5	`claude-opus-4-5-20251101`	not compared	Retirement not sooner than November 24, 2026
Claude Opus 5.2	none	not released	No such model

Opus 4.1 and Opus 4 are retired. Release dates are given only for the two Opus 5 models; the older rows come from Anthropic's deprecations table.

Opus 5.5 specs: the model to use instead

Published figures for the current Opus as of November 2026.

Maker: Anthropic
Released: September 22, 2026
Claude API ID: claude-opus-5-5
Amazon Bedrock ID: anthropic.claude-opus-5-5
Input / output: Text and images in, text out
Context window: 1M tokens
Max output: 128K tokens; 300K on the Batch API with a beta header
Thinking: Adaptive, always on; depth set by the effort parameter
Default effort (API): medium
Knowledge cutoff: June 2026
Retirement: Not sooner than September 22, 2027

How much does the latest Claude Opus cost?

Opus 5.5 costs $4 per million input tokens and $20 per million output tokens, 20% less per token than Opus 5. Fable 5.2 and Mythos 5.2 share one price. Prices as of November 2026.

Input, per MTok

Opus 5.5: $4
Opus 5: $5
Fable / Mythos 5.2: $8

Output, per MTok

Opus 5.5: $20
Opus 5: $25
Fable / Mythos 5.2: $40

5-minute cache write

Opus 5.5: $5
Opus 5: $6.25
Fable / Mythos 5.2: $10

1-hour cache write

Opus 5.5: $8
Opus 5: $10
Fable / Mythos 5.2: $16

Cache read

Opus 5.5: $0.20
Opus 5: $0.50
Fable / Mythos 5.2: $0.20

Batch API (50% off)

Opus 5.5: $2 / $10
Opus 5: $2.50 / $12.50
Fable / Mythos 5.2: $4 / $20

Fast mode

Opus 5.5: $8 / $40
Opus 5: not compared
Fable / Mythos 5.2: $16 / $80

Example: 100K in + 10K out

Opus 5.5: $0.60
Opus 5: $0.75
Fable / Mythos 5.2: $1.20

Rate	Opus 5.5	Opus 5	Fable / Mythos 5.2
Input, per MTok	$4	$5	$8
Output, per MTok	$20	$25	$40
5-minute cache write	$5	$6.25	$10
1-hour cache write	$8	$10	$16
Cache read	$0.20	$0.50	$0.20
Batch API (50% off)	$2 / $10	$2.50 / $12.50	$4 / $20
Fast mode	$8 / $40	not compared	$16 / $80
Example: 100K in + 10K out	$0.60	$0.75	$1.20

The example row is our arithmetic from the list prices: 0.1M × $4 plus 0.01M × $20 = $0.60 for Opus 5.5. Batch rows are half the list price, per Anthropic's 50% batch discount.

How does Opus 5.5 score against Opus 5?

Scores Anthropic published at the Opus 5.5 launch, at max effort unless noted. We have not run them ourselves.

Terminal-Bench 4.0 (agentic coding)

Opus 5.5: 66.4%
Fable 5.1: 55.8%
Opus 5: 52.3%

FrontierCode v1.1, main set

Opus 5.5: 54.4%
Fable 5.1: 50.3%
Opus 5: 48.0%

CursorBench 4.0

Opus 5.5: 57.8%
Fable 5.1: 51.8%
Opus 5: 46.6%

GDPval-AA v2.1 (knowledge work)

Opus 5.5: 1846 Elo
Fable 5.1: 1735 Elo
Opus 5: 1708 Elo

Humanity's Last Exam (with tools)

Opus 5.5: 67.7%
Fable 5.1: 65.6%
Opus 5: 63.6%

OSWorld 2.1 (computer use, partial)

Opus 5.5: 81.8%
Fable 5.1: 80.7%
Opus 5: 74.0%

Benchmark	Opus 5.5	Fable 5.1	Opus 5
Terminal-Bench 4.0 (agentic coding)	66.4%	55.8%	52.3%
FrontierCode v1.1, main set	54.4%	50.3%	48.0%
CursorBench 4.0	57.8%	51.8%	46.6%
GDPval-AA v2.1 (knowledge work)	1846 Elo	1735 Elo	1708 Elo
Humanity's Last Exam (with tools)	67.7%	65.6%	63.6%
OSWorld 2.1 (computer use, partial)	81.8%	80.7%	74.0%

Vendor-reported. Opus 5.5's Terminal-Bench score is at xhigh effort. Anthropic itself says benchmark margins at this level are a less reliable guide to real-world differences.

What Opus 5.5 changed, for anyone expecting Claude Opus 5.2

If you were waiting for a Claude Opus 5.2 upgrade, these are the improvements that shipped in Opus 5.5 instead.

Cheaper to run

Token prices fell 20% from Opus 5, cache reads 60% (from $0.50 to $0.20 per million), and Anthropic's tests put the total at 40% less on typical workloads because it also uses fewer tokens.

Faster output

Output is more than 30% faster than Opus 5. An optional fast mode runs up to 2.5 times faster at $8 / $40 per million tokens.

Fable 5.1-level work

Anthropic says Opus 5.5 performs at the level of Fable 5.1 on most work. In one internal test it ported HAProxy from C to Rust in 9.5 hours against 12 for Fable 5.1, at 51% less cost.

Stronger safety scores

It posted the best result to date on Anthropic's automated behavioral audit and is more resistant than Opus 5 to prompt injection.

Start of the 5.5 family

Opus 5.5 was the first Claude 5.5 model. Sonnet 5.5 and Haiku 5.5 followed, and all three are on the current lineup.

Clearer writing

Testers found its writing clearer and easier to follow than Opus 5, with the most important information first.

How to access the current Opus

Opus 5.5 is generally available, with no waitlist or verification step.

Claude apps

The Free plan does not include Opus. Pro ($20 a month, or $17 billed yearly), Max ($100) and Ultra ($200) all include Opus models.

Claude API

Call the Messages API with claude-opus-5-5, billed per token at the rates above. Batch requests get 50% off, and fast mode costs double the standard rate.

Cloud platforms

Opus 5.5 also runs on Amazon Bedrock (anthropic.claude-opus-5-5), Claude Platform on AWS, Claude Models on Google Cloud and Microsoft Foundry, each billed through its own account.

Opus 5.5 API gotchas before you migrate

Some request settings that worked on Opus 5 now return errors.

Control thinking depth with the effort parameter; the default is medium
Use tool_choice set to auto with strict tool use
Switch computer use to the computer_toolset_20260801 toolset
Pin the full model ID claude-opus-5-5 in production

Don't

Send thinking: disabled or a manual thinking budget; both return a 400
Force tools with tool_choice of any or tool
Request a model called claude-opus-5-2; it does not exist
Assume Opus 5.5 reads thinking blocks written by Fable 5.2; the API drops them

Does Synexa run Claude Opus?

No. Synexa does not serve Opus 5.5, Fable, Mythos or any other Claude model, and it offers no chat LLMs; its only text outputs are LLaVA-13B image answers, Whisper speech transcripts and an NSFW image-check label. Use the Claude API or one of the clouds above for the language side.

Synexa covers the media half of an AI product through one REST API. One key gives you image models such as Nano Banana Pro at $0.10 a request, music, speech, upscaling, and image-to-3D models such as Hunyuan3D 3.1 at $0.16 a request. A common pattern is to have Opus write the prompt, then send it to Synexa to make the asset. Every price is listed in Explore, new accounts get $0.10 of free credit, and the quickstart has Python and Node examples.

Pay per request, no subscription
Python and Node clients, plus plain HTTP
Image, video, audio, upscaling and 3D in one catalog

Synexa models to pair with an Opus agent

Media an Opus-driven agent can request from Synexa. Flat prices per request, as listed on each model page in November 2026.

Product photos and ad visuals

Synexa model: Nano Banana Pro
Price per request: $0.10

Posters with readable text

Synexa model: GPT Image 2
Price per request: $0.10

Textured 3D model from one image

Synexa model: Hunyuan3D 3.1
Price per request: $0.16

Draft 3D mesh

Synexa model: TRELLIS
Price per request: $0.02

Background music clip

Synexa model: Lyria 2
Price per request: $0.10

Meeting or video transcript

Synexa model: Whisper v3
Price per request: $0.004

Task	Synexa model	Price per request
Product photos and ad visuals	Nano Banana Pro	$0.10
Posters with readable text	GPT Image 2	$0.10
Textured 3D model from one image	Hunyuan3D 3.1	$0.16
Draft 3D mesh	TRELLIS	$0.02
Background music clip	Lyria 2	$0.10
Meeting or video transcript	Whisper v3	$0.004

Synexa catalog prices, not Anthropic's. Synexa does not run any Claude model.

Should you wait for Claude Opus 5.2 or use Opus 5.5 now?

Use Opus 5.5 now. Anthropic has given no sign of a Claude Opus 5.2, and its own guidance for the November 9 launch is to start most coding and knowledge work on Opus 5.5, which it calls its strongest daily driver, then move up to Fable 5.2 or Mythos 5.2 only when the task is long, open-ended or still failing.

The step up costs twice as much per token ($8 / $40 against $4 / $20). Anthropic reports that Fable 5.2 at high effort scores about level with Opus 5.5 at max effort on FrontierCode 2.0 for roughly half the cost per task, so the upgrade can pay off on long agent runs. Step down to Sonnet 5.5 or Haiku 5.5 for chat, extraction and routing. If your product also generates images, voice or 3D, compare media APIs in our Runware alternative and WaveSpeed AI alternative pages.

The Claude 5.2 models and more API comparisons.

Claude Mythos 5.2

Anthropic's verification-only twin of Fable 5.2: who can use it, what it costs, and how to apply.

Claude Fable 5.2

Anthropic's most capable generally available model: specs, pricing, access and API gotchas.

Synexa vs Runware

Synexa compared with Runware: per-model prices, free credit, catalog size and when each one fits.

Explore Synexa models

Browse the image, video, audio and 3D models you can run with one API key.

Synexa pricing

Per-request prices for the models on Synexa, with no subscription.

Frequently asked questions

Is Claude Opus 5.2 out?

No. As of November 14, 2026, Anthropic has not released a model called Claude Opus 5.2, and its documentation, model list and release notes contain no such model. The latest Opus is Opus 5.5, released September 22, 2026. The models numbered 5.2 are Claude Fable 5.2 and Claude Mythos 5.2, both released on November 9, 2026.

What is the latest Claude Opus model?

Claude Opus 5.5, with the API ID claude-opus-5-5, released September 22, 2026. It has a 1M-token context window, returns up to 128K tokens, and costs $4 per million input tokens and $20 per million output tokens. Anthropic marks it as the latest Opus and recommends it as the starting point for most coding and knowledge work.

Did Anthropic skip Opus 5.1 and 5.2?

Yes, for Opus. The Anthropic Opus lineup goes from Opus 5 in July 2026 to Opus 5.5 in September 2026, with no Opus 5.1 or 5.2 in its model list or release notes. The 5.1 and 5.2 numbers were used for the Fable and Mythos models instead, which sit above Opus in capability and price.

How much does Opus 5.5 cost compared with Opus 5?

Opus 5.5 lists at $4 per million input tokens and $20 per million output tokens, against $5 and $25 for Opus 5. Cache reads dropped from $0.50 to $0.20 per million. Anthropic says the combined effect, including fewer tokens per task, is about 40% lower cost on typical workloads. A call with 100K tokens in and 10K out costs $0.60.

Is Opus 5.5 better than Fable 5.2?

No, Fable 5.2 is the stronger model. Anthropic's launch table shows Fable 5.2 ahead of Opus 5.5 on almost every benchmark it reported, such as 69.9% against 62.3% on FrontierSWE v2. Fable also costs twice as much per token, so Anthropic suggests starting on Opus 5.5 and moving to Fable only for the hardest, longest tasks.

Can I use Opus on the free Claude plan?

No. The Claude Free plan includes Sonnet and Haiku but not Opus. Opus models are included in Pro ($20 a month, or $17 billed yearly), Max ($100 a month) and Ultra ($200 a month). On the API, Opus is billed per token from a funded Claude Platform or cloud account.

What model ID should I use instead of claude-opus-5-2?

Use claude-opus-5-5 on the Claude API, Google Cloud, Microsoft Foundry and Claude Platform on AWS, or anthropic.claude-opus-5-5 on Amazon Bedrock. A request for claude-opus-5-2 will fail because no such model exists. If you need more capability than Opus 5.5, the next step up is claude-fable-5-2.

Can I use Claude Opus through Synexa?

No. Synexa does not serve Claude Opus or any other chat language model. Synexa is a pay-per-request API for generative media: images, video, speech and music, upscaling and image-to-3D. Many teams call Claude for planning and writing, then call Synexa to render the image, clip, voice-over or 3D model. New Synexa accounts get $0.10 of free credit.

Pair Opus with a media API

Use the latest Claude Opus for reasoning and writing, then generate the images, video, music and 3D models it plans on Synexa. Pay per request, start with $0.10 free credit.