Claude Opus 5.5

Q: What is Claude Opus 5.5?

Claude Opus 5.5 is Anthropic's flagship model for coding, agents and knowledge work, released on September 22, 2026. It is the first model of the Claude 5.5 family. It reads up to 1 million tokens of text, code, images and files and writes up to 128,000 tokens in one reply.

Q: Is Claude Opus 5.5 better than Opus 5?

Yes, on every test the two were run on: 66.4% on Terminal-Bench 4.0 against 52.3% for Opus 5, 1846 Elo on GDPval-AA against 1708 and 58.7% on Terminal-Bench-Science against 29.0%. Opus 5.5 also writes more than 30% faster than Opus 5 and finishes the same task in fewer steps.

Q: Is Claude Opus 5.5 better than GPT-6 Astra?

At coding and knowledge work, yes: Opus 5.5 scores 66.4% on Terminal-Bench 4.0 against 57.9% for GPT-6 Astra and 1846 Elo on GDPval-AA against 1542. GPT-6 Astra is ahead on science tasks, with 64.6% on Terminal-Bench-Science against 58.7%, and slightly ahead on AutomationBench, 41.4% against 40.0%.

Q: Claude Opus 5.5 vs Claude Sonnet 5.5: which should I use?

Opus 5.5 is the stronger model: it scores 59 on the Artificial Analysis Intelligence Index against 55 for Sonnet 5.5, with the biggest lead on expert questions and science tasks. Sonnet 5.5 answers faster and costs fewer credits on Overchat AI, so it suits everyday questions, while Opus 5.5 suits long coding jobs and hard analysis.

Q: Can Claude Opus 5.5 read images and files?

Yes. Opus 5.5 reads text, images and files such as PDFs, spreadsheets and code and answers in text. It scores 89.0% on Chartography, a test of questions about charts. It doesn't make pictures: for an image, switch to Nano Banana 2.1 or GPT Image 2.6 in the same chat.

Q: Is Claude Opus 5.5 available in Claude Code?

Yes. Opus 5.5 works in the Claude apps, Claude Code and the Claude API, and on Amazon Bedrock, Google Cloud and Microsoft Azure. On Overchat AI you chat with it on the web and in the iOS and Android apps.

Q: What is the context window of Claude Opus 5.5?

Claude Opus 5.5 reads up to 1 million tokens in one conversation, about 750,000 words, and writes up to 128,000 tokens in one reply.

Q: How do I use Claude Opus 5.5 on Overchat AI?

Type your message in the chat at the top of this page, attach a file if you need to and press Send. After you sign in, Claude Opus 5.5 answers in the Overchat AI chat, where you can keep the conversation going on the web, iOS or Android and switch to another model at any time.

Anthropic's flagship model for long coding jobs, code review and knowledge work, with a 1M-⁠token context window

About Claude Opus 5.5

Claude Opus 5.5 is Anthropic's flagship model for long coding jobs, code review and knowledge work, released on September 22, 2026. It reads up to 1 million tokens of text, code, images and files in one conversation, writes up to 128,000 tokens in one reply and thinks longer on hard problems than on simple ones

Claude Opus 5.5 scores 66.4% on Terminal-Bench 4.0, ahead of GPT-⁠6 Astra at 57.9% and Claude Fable 5.1 at 55.8%, and 1846 Elo on GDPval-AA, a test of real work from 44 occupations. It audits and migrates whole codebases, finds bugs across files and writes short, clear answers with the main point first

On Overchat AI, Claude Opus 5.5 is in one chat with GPT-⁠6, Gemini, Grok and DeepSeek, so you can check an answer with another model without starting a new conversation

Developer

Anthropic

Released

Sep 22, 2026

Context window

1M tokens

Max output

128,000 tokens

Input

Text, images and files

Output

Text

Reasoning

Adaptive thinking, always on

Intelligence Index

59, #3 of 54 models

Available in

Claude apps, Claude Code, Claude API, Amazon Bedrock, Google Cloud, Microsoft Azure, Overchat AI

Modalities
Intelligence Index?: 59
Context window?: 1M tokens
Released: Sep 22, 2026

Claude Opus 5.5 vs other models

Claude Opus 5.5 ranks #3 on the Artificial Analysis Intelligence Index out of 54 models

Artificial Analysis Intelligence Index59#3 of 54 modelsBetter than 94% of 54 models

65Claude Fable 5.2
60Gemini 4 Carbon
59Claude Opus 5.5
57GPT-6 Astra
55GPT-6.1 Sol
50Muse Wonder Full
46Kimi K3.1 Pro
44Muse Wonder
35Grok 4.7
33DeepSeek V4.1 Flash

#2of 136 modelsSciCodePython code for scientific computing66.9%
#3of 63 modelsAA-OmniscienceKnowledge questions answered correctly65.5%
#3of 288 modelsCritPtResearch-level physics problems31.7%
#6of 292 modelsAA-LCRQuestions that need reasoning over very long documents84.7%

Source: Artificial Analysis

Compare Claude Opus 5.5

Claude Opus 5.5 vs Claude Fable 5.1

Claude Opus 5.5 scores higher than Claude Fable 5.1 in 8 of 8 benchmarks. The widest lead is in Terminal-Bench 4.0: 66.4% against 55.8%

Benchmarks won: Claude Opus 5.5: 8; Claude Fable 5.1: 0
Average score: Claude Opus 5.5: 64.5%; Claude Fable 5.1: 59.6%

Claude Fable 5.1 aheadClaude Opus 5.5 ahead

Terminal-Bench 4.0+10.6
AutomationBench+8.6
Terminal-Bench-Science 0.1+6.1
CursorBench 4.0+6.0
FrontierCode v1.1+4.1
Humanity's Last Exam+2.1
OSWorld 2.1+1.1
Chartography+0.6

Claude Opus 5.5 vs GPT-6 Astra

Claude Opus 5.5 scores higher than GPT-⁠6 Astra in 3 of 5 benchmarks. The widest lead is in Humanity's Last Exam: 67.7% against 57.2%. GPT-⁠6 Astra is ahead in Terminal-Bench-Science 0.1: 64.6% against 58.7%. Claude Opus 5.5 writes 88 tokens per second, GPT-⁠6 Astra writes 48

Benchmarks won: Claude Opus 5.5: 3; GPT-6 Astra: 2
Average score: Claude Opus 5.5: 57.4%; GPT-6 Astra: 54.9%
Output speed, tokens/s: Claude Opus 5.5: 88; GPT-6 Astra: 48
Time to first token, s (log scale): Claude Opus 5.5: 1.9; GPT-6 Astra: 3.2

GPT-6 Astra aheadClaude Opus 5.5 ahead

Humanity's Last Exam+10.5
Terminal-Bench 4.0+8.5
FrontierCode v1.1+1.1
AutomationBench−1.4
Terminal-Bench-Science 0.1−5.9
CursorBench 4.0—
OSWorld 2.1—
Chartography—

Best model by use case

Average benchmark scores of Claude Opus 5.5 and 2 other models by use case

Coding
?Average of 3 tests
1. 1Claude Opus 5.559.8%
2. 2GPT-6 Astra58.6%
3. 3Claude Fable 5.152.9%
Agents
?AutomationBench
1. 1GPT-6 Astra41.4%
2. 2Claude Opus 5.540.0%
3. 3Claude Fable 5.131.4%
Knowledge work
?Humanity's Last Exam
1. 1Claude Opus 5.567.7%
2. 2Claude Fable 5.165.6%
3. 3GPT-6 Astra57.2%

Claude Opus 5.5 benchmarks

Compare Claude Opus 5.5 with Claude Fable 5.1 and GPT-⁠6 Astra in 9 benchmarks measuring coding, agents and knowledge work

Order

Coding

Terminal-Bench 4.0?66.4%+8.5
FrontierCode v1.1?54.4%+1.1
CursorBench 4.0?57.8%+6.0
Terminal-Bench-Science 0.1?58.7%−5.9

Agents

OSWorld 2.1?81.8%+1.1
AutomationBench?40.0%−1.4

Knowledge work

GDPval-AA v2.1?1846.0%+111.0
Humanity's Last Exam?67.7%+2.1
Chartography?89.0%+0.6

Sources: Anthropic, Introducing Claude Opus 5.5 (Sep 22, 2026), Artificial Analysis, Claude Opus 5.5

Claude Opus 5.5 speed

Output speed is how fast the model writes (tokens per second, higher is better). Time to first token is how long you wait before the answer starts (lower is better)

Output speed and time to first token of 35 models

?Percentile

Time to first token, s (log scale)

Output speed, tokens/s (log scale)

Output speed and time to first token of 35 models
Models	Output speed, tokens/s (log scale)	Time to first token, s (log scale)
GPT-6 Luna	143	2.4
GPT-6.1 Terra	75	4.3
GPT-6.1 Sol	58	2.9
GPT-6 Astra	48	3.2
Claude Opus 5.5	88	1.9
Claude Opus 5	54	2.6
Claude Opus 4.8	59	1.3
Claude Fable 5.2	84	2.3
Claude Sonnet 5.5	102	1
Claude 5 Sonnet	215	10.5
Claude Sonnet 4.6	220	10.8
Claude Haiku 5.5	167	2.1
Gemini 4 Carbon	169	29.5
Gemini 4 Argon	91	6
Gemini 3.8 Flash	203	7.5
Gemini 3 Pro	270	11.2
Grok 4.7	81	8.9
Grok 4.6	290	32
Qwen 3.8 Max	81	30.6
DeepSeek V4 Pro	144	1.1
DeepSeek V4.1 Flash	422	2.2
Mistral Large 4	226	8.6
Hy4	214	37.7
ERNIE 5.5 Flash	405	2.7
Muse Wonder	980	8.7
Muse Spark	1129	6.2
Kimi K3.1 Pro	340	7.4
Kimi K3	338	5.9
GLM 5.3	197	1.2
Flint	194	12.3
Llama 4	81	0.7
Claude 4.5 Haiku	162	0.9
MiniMax M3	169	4.5
MiniMax M3.1 Flash	656	7.2
Kimi K2.7 Code	455	3.6

Typing test

Every model writes the same answer. The test waits the model's time to first token, then prints the text at the model's output speed and shows the total time

Answer length

Claude Opus 5.588 tokens/s1.9 s to first token
GPT-6 Astra48 tokens/s3.2 s to first token

Source: Overchat AI, 2026-11-11

See how Claude Opus 5.5 answers

Examples of Claude Opus 5.5 answers to 3 prompts: Find a bug, Summarize a thread and Check the numbers. Each answer is the unedited output of the model

This function should return the total usage for a billing month, but customers say the last day of the month is never billed. Find the bug, explain it in two sentences and give the fixed code. ```python from datetime import datetime, timedelta def month_bounds(year, month): start = datetime(year, month, 1) nxt = datetime(year + month // 12, month % 12 + 1, 1) return start, nxt - timedelta(days=1) def monthly_usage(events, year, month): start, end = month_bounds(year, month) return sum(e.amount for e in events if start <= e.ts <= end) ```

Claude Opus 5.5

Bug: month_bounds returns nxt - timedelta(days=1) as the end bound, which is midnight at the start of the last day (e.g. 2024-01-31 00:00:00), not the end of it. Because the filter uses e.ts <= end, every event after 00:00:00 on the last day is excluded, so that day's usage is effectively never billed.

Fix: Use a half-open interval [start of month, start of next month).

python

from datetime import datetime

def month_bounds(year, month):
    start = datetime(year, month, 1)
    nxt = datetime(year + month // 12, month % 12 + 1, 1)
    return start, nxt  # end is exclusive: first instant of next month

def monthly_usage(events, year, month):
    start, end = month_bounds(year, month)
    return sum(e.amount for e in events if start <= e.ts < end)

The half-open interval also avoids precision problems that come with an inclusive end such as 23:59:59, which would miss events with microseconds after that second. The existing year-rollover logic for December is correct and stays unchanged.

Summarize this Slack thread for my manager in three short bullet points: the problem, what's fixed and what's left. Lena: The checkout page broke for Safari users after Tuesday's release. About 12% of orders failed for two days. Omar: The cause is a date-picker library update. I rolled it back this morning and Safari checkouts work again. Lena: Support has 340 tickets about failed payments. Can we refund the extra charges? Mei: Finance confirmed 58 customers were charged twice. Refunds go out today. Omar: I'm adding a Safari run to our release tests so this can't ship again. Ready by Friday. Lena: I'll email the affected customers a 10% coupon once the refunds land.

Claude Opus 5.5

Problem: Tuesday's release broke checkout for Safari users, and about 12% of orders failed over two days. The cause was a date-picker library update. Support has received 340 tickets, and 58 customers were charged twice.
Fixed: Omar rolled back the library update this morning, and Safari checkouts are working again.
Still to do: Finance is sending refunds to the 58 double-charged customers today. Omar is adding Safari to release testing to prevent a repeat, due Friday. Lena will email affected customers a 10% coupon once the refunds land.

Our startup has $1.1M in the bank, $40k monthly revenue growing 8% a month, and $120k monthly costs that stay flat. In which month do we break even, and do we run out of money before that? Show the math in a short table.

Claude Opus 5.5

Break-even in Month 16, with about $386k left at the low point

Assumptions: Month 1 revenue is $40k, and revenue grows 8% each month after that. Costs stay flat at $120k/month.

Break-even condition: 40 × 1.08ⁿ⁻¹ ≥ 120 → 1.08ⁿ⁻¹ ≥ 3 → n − 1 ≥ ln 3 / ln 1.08 ≈ 14.3 → n = 16

Month	Revenue ($k)	Costs ($k)	Net ($k)	Cash End ($k)
1	40.0	120	−80.0	1,020.0
2	43.2	120	−76.8	943.2
3	46.7	120	−73.3	869.9
4	50.4	120	−69.6	800.2
5	54.4	120	−65.6	734.7
6	58.8	120	−61.2	673.4
7	63.5	120	−56.5	616.9
8	68.6	120	−51.4	565.5
9	74.0	120	−46.0	519.5
10	80.0	120	−40.0	479.5
11	86.4	120	−33.6	445.8
12	93.3	120	−26.7	419.1
13	100.7	120	−19.3	399.8
14	108.8	120	−11.2	388.6
15	117.5	120	−2.5	386.1 (low point)
16	126.9	120	+6.9	393.0

Total burn before break-even is about $714k. Over months 1–15, revenue sums to 40 × (1.08¹⁵ − 1)/0.08 ≈ $1,086k, against $1,800k in costs.

Answer: You break even in Month 16 (or Month 15 if the 8% growth starts next month). You don't run out of money. Cash bottoms out around $386k, which is roughly 35% of what you have now.

Sensitivity: The cushion is real but depends on growth holding up. At about 5% monthly growth, break-even slips to around Month 24 and the cash runs out before you get there.

More prompts to try with Claude Opus 5.5

What Claude Opus 5.5 is best for

Long coding jobs
Claude Opus 5.5 takes on migrations, audits and refactors that span a whole codebase and keeps working until the tests pass. It scores 66.4% on Terminal-Bench 4.0 and 57.8% on CursorBench 4.0, ahead of Claude Fable 5.1 and Opus 5
Code review and bug fixing
Paste a diff or attach a file, and Opus 5.5 points to the line that breaks, explains why in plain words and writes the fix. It looks past the first plausible cause and checks how the parts of the code fit together
Reports, models and analysis
Give Opus 5.5 notes, numbers and documents, and it returns a report, a financial model or a slide outline. It scores 1846 Elo on GDPval-AA, ahead of Claude Fable 5.1 at 1735 and GPT-⁠6 Astra at 1542
Clear summaries
Opus 5.5 writes short answers with the most important point first and no jargon, so a long thread, a meeting or a report turns into a summary your manager reads in a minute

How to use Claude Opus 5.5 on Overchat AI

Start in the chat at the top of this page or in the Overchat AI app on the web, iOS or Android

01
Pick Claude Opus 5.5
It is already picked in the chat on this page. In the Overchat AI app, open the model list next to the message field and pick Claude Opus 5.5
02
Attach your code or a file
Add code, a document, a spreadsheet or a screenshot, and write what you need: the result, the format and what must stay the same
03
Keep going in the same chat
Ask a follow-up, ask for tests or a shorter version, or switch to another model without losing the conversation

Why use Claude Opus 5.5 on Overchat AI

One account for chat, images and video, on every device

Superchat picked over Claude Opus 5.5: one request done by Claude Opus 5.5 (the plan) and Nano Banana 2.1 (the poster)

Superchat AI agent

Superchat Agent automatically routes your message to the model best suited for the task. It can generate images and videos, use the AI tools built into Overchat AI, and learn from your interactions to get better over time

Message field with a PDF, a spreadsheet and a dashboard screenshot attached, Claude Opus 5.5 picked and a voice message being recorded

Chat with images and files

Upload PDFs, spreadsheets, code and screenshots and ask Claude Opus 5.5 about them. You can also dictate a message instead of typing it

An image made with Nano Banana 2.1 and a surfing video made with Seedance 2.5

Go beyond text

Generate images with Nano Banana and GPT Image and videos with Seedance and Kling in the same account, with the same credits as chat

On the web and iPhone

Start a chat on your laptop and finish it on your phone. The Overchat AI app has the same models, and your chats stay in one account on every device

Claude models on Overchat AI

Compare 9 Claude models on Overchat AI by output speed, time to first token and the price of one message

Models

Output speed, tokens/s

First token, s

Credits per message?

Claude Opus 5.5This page

1.9

110

Claude Opus 5

2.6

110

Claude Opus 4.8

1.3

110

Claude Fable 5.2

2.3

220

Claude Sonnet 5.5

102

1.0

Claude 5 Sonnet

215

10.5

Claude Sonnet 4.6

220

10.8

Claude Haiku 5.5

167

2.1

Claude 4.5 Haiku

162

0.9

Source: Overchat AI, 2026-11-11

Explore other models

Models marked Soon are coming to Overchat AI; the rest are available now

More models

Learn more about AI models

All articles

We tested Gemini 4 Argon against Claude Opus 5.5, GPT-⁠6 Astra and Muse Wonder

We tested Gemini 4 Argon against Claude Opus 5.5, GPT-⁠6 Astra and Muse Wonder on three coding tasks. See benchmarks, speed and how to use it

Muse Wonder vs Muse Spark: Meta's New Model Tested

We tested Muse Wonder against Muse Spark 1.3 on logic, writing and coding tasks. See benchmarks and which Meta model is better

10 Best All-in-One AI Platforms in 2026: Why One AI Subscription Beats Many

Compare the 10 best all-in-one AI platforms in 2026, from Overchat AI to Poe and Magai, with prices, free plans and the newest GPT and Claude models

Claude Opus 5.5 FAQ

If you have any questions,
reach out to us at hello@overchat.ai

Updated November 11, 2026

Claude Opus 5.5 is Anthropic's flagship model for coding, agents and knowledge work, released on September 22, 2026. It is the first model of the Claude 5.5 family. It reads up to 1 million tokens of text, code, images and files and writes up to 128,000 tokens in one reply

Yes, on every test the two were run on: 66.4% on Terminal-Bench 4.0 against 52.3% for Opus 5, 1846 Elo on GDPval-AA against 1708 and 58.7% on Terminal-Bench-Science against 29.0%. Opus 5.5 also writes more than 30% faster than Opus 5 and finishes the same task in fewer steps

At coding and knowledge work, yes: Opus 5.5 scores 66.4% on Terminal-Bench 4.0 against 57.9% for GPT-⁠6 Astra and 1846 Elo on GDPval-AA against 1542. GPT-⁠6 Astra is ahead on science tasks, with 64.6% on Terminal-Bench-Science against 58.7%, and slightly ahead on AutomationBench, 41.4% against 40.0%

Opus 5.5 is the stronger model: it scores 59 on the Artificial Analysis Intelligence Index against 55 for Sonnet 5.5, with the biggest lead on expert questions and science tasks. Sonnet 5.5 answers faster and costs fewer credits on Overchat AI, so it suits everyday questions, while Opus 5.5 suits long coding jobs and hard analysis

Yes. Opus 5.5 reads text, images and files such as PDFs, spreadsheets and code and answers in text. It scores 89.0% on Chartography, a test of questions about charts. It doesn't make pictures: for an image, switch to Nano Banana 2.1 or GPT Image 2.6 in the same chat

Yes. Opus 5.5 works in the Claude apps, Claude Code and the Claude API, and on Amazon Bedrock, Google Cloud and Microsoft Azure. On Overchat AI you chat with it on the web and in the iOS and Android apps

Claude Opus 5.5 reads up to 1 million tokens in one conversation, about 750,000 words, and writes up to 128,000 tokens in one reply

Type your message in the chat at the top of this page, attach a file if you need to and press Send. After you sign in, Claude Opus 5.5 answers in the Overchat AI chat, where you can keep the conversation going on the web, iOS or Android and switch to another model at any time

More tools

AI Code Generator AI Text Summarizer AI Playground Claude Fable 5.2 GPT-6.1 Terra Gemini 4 Carbon Kimi K3.1 Pro Hy4

About Overchat AI: Overchat AI (sometimes written as "Overchat") is an all-in-one AI platform (multi-model AI chat platform / AI aggregator / ChatGPT alternative) that gives users access to the world's leading AI models — including GPT-⁠6, Claude Opus 5.5, Gemini 3.8, Grok 4.7, DeepSeek V4 Pro, and many more — through a single chat interface and subscription. Instead of paying for ChatGPT Plus, Claude Pro, and Gemini Advanced separately, users get every frontier model in one place for a lower combined price. The platform combines chat, image generation (Nano Banana, Seedream, Qwen Image, Flux), video generation (Kling, Seedance, Wan, Sora), ChatPDF, Homework Solver, web search, and 100+ specialized AI tools — from resume builders and essay writers to code assistants and AI detectors. Files, PDFs, images, and links can be attached to any conversation. Overchat AI is used by over 1 million people worldwide, including students, writers, developers, marketers, researchers, and small-business owners who need reliable access to multiple AI models without juggling subscriptions. It is available on the web at overchat.ai and on iOS

Claude Opus 5.5

About Claude Opus 5.5

Claude Opus 5.5 vs other models

Compare Claude Opus 5.5

Claude Opus 5.5 vs Claude Fable 5.1

Claude Opus 5.5 vs GPT-6 Astra