Anthropic is releasing Claude Opus 5.5, the opening model in its Claude 5.5 family and the current top of the Opus tier aimed at engineers, agent builders, and enterprise teams doing serious knowledge work. Claude Opus 5.5 is Anthropic's strongest Opus model yet, powering long-running, highly capable agents while delivering improvements in coding and professional work . The launch matters because it collapses the price gap between Opus-tier intelligence and cheaper models, making frontier reasoning practical for workloads that used to be too expensive to run at Opus quality.
The pitch centers on cost and reliability rather than raw capability jumps. Opus 5.5 costs $4 per million input tokens and $20 per million output tokens, 20% below Opus 5 , and cache reads, which are a large share of the cost of long-running agentic work, also now cost 60% less than Opus 5, at $0.20 per million tokens . For business users and consumers who want to collaborate with a powerful model on complex tasks, Opus 5.5 is available on Claude for Pro, Max, Team, and Enterprise users , and it ships on the Claude API alongside AWS, Google Cloud, and Microsoft Foundry.
Under the hood, the model is positioned for agentic coding across large codebases, multi-tool orchestration with subagents and memory, financial analysis on dense filings, and computer use across multiple applications. Anthropic was founded in 2021 by a group of former OpenAI researchers, including Dario Amodei, Daniela Amodei, Jared Kaplan, and Chris Olah. The company is a privately held public benefit corporation , and Opus 5.5 continues its pattern of pairing model releases with a system card and updated safeguards in areas like cybersecurity and biology.
Comments (7)
Sign in to join the discussion.
P
Priya Vasquez21d ago
40% cheaper than Opus 5 but performs like Fable 5.1? So you're telling me the previous flagship is now the mid-tier benchmark. My finance brain is doing gymnastics on the depreciation curve here.
S
solange.eth21d ago
imagine if compute credits for this were onchain and tradeable. the arb between Opus 5 and 5.5 pricing alone would be a whole desk.
K
kwesi21d ago
The naming is starting to feel like a Toyota trim lineup. Opus, Fable, Sonnet, next week we get Claude Corolla LE.
M
Mikaela R.21d ago
ok wait, dropping a new flagship on a Tuesday with a two-sentence tweet and no thread is such a power move. no launch video, no CEO essay, just numbers go up cost go down.
R
Renzo Koba21d ago
Keep thinking is a great tagline until my agent loop keeps thinking for 40 minutes on a task that should take 4. reliability numbers or it didn't happen.
T
Tavita Ng21d ago
Interesting. Would love to see retention on Opus 5 users actually migrating vs staying pinned to the old version for reproducibility. Cheaper doesn't matter if nobody switches.
B
Bogdan Maros21d ago
The screenshot in the tweet has that classic Anthropic beige energy. At this point the muted palette is doing more brand work than the model names are.
Work With Us
Book an intro call to explore our suite of services. We might have a waitlist.
Anthropic
Keep thinking.
Keep thinking.
Tags
TLVC Rating
Community Rating
No ratings yet
to rate this launch.
Anthropic is releasing Claude Opus 5.5, the opening model in its Claude 5.5 family and the current top of the Opus tier aimed at engineers, agent builders, and enterprise teams doing serious knowledge work. Claude Opus 5.5 is Anthropic's strongest Opus model yet, powering long-running, highly capable agents while delivering improvements in coding and professional work . The launch matters because it collapses the price gap between Opus-tier intelligence and cheaper models, making frontier reasoning practical for workloads that used to be too expensive to run at Opus quality. The pitch centers on cost and reliability rather than raw capability jumps. Opus 5.5 costs $4 per million input tokens and $20 per million output tokens, 20% below Opus 5 , and cache reads, which are a large share of the cost of long-running agentic work, also now cost 60% less than Opus 5, at $0.20 per million tokens . For business users and consumers who want to collaborate with a powerful model on complex tasks, Opus 5.5 is available on Claude for Pro, Max, Team, and Enterprise users , and it ships on the Claude API alongside AWS, Google Cloud, and Microsoft Foundry. Under the hood, the model is positioned for agentic coding across large codebases, multi-tool orchestration with subagents and memory, financial analysis on dense filings, and computer use across multiple applications. Anthropic was founded in 2021 by a group of former OpenAI researchers, including Dario Amodei, Daniela Amodei, Jared Kaplan, and Chris Olah. The company is a privately held public benefit corporation , and Opus 5.5 continues its pattern of pairing model releases with a system card and updated safeguards in areas like cybersecurity and biology.
Comments (7)
40% cheaper than Opus 5 but performs like Fable 5.1? So you're telling me the previous flagship is now the mid-tier benchmark. My finance brain is doing gymnastics on the depreciation curve here.
imagine if compute credits for this were onchain and tradeable. the arb between Opus 5 and 5.5 pricing alone would be a whole desk.
The naming is starting to feel like a Toyota trim lineup. Opus, Fable, Sonnet, next week we get Claude Corolla LE.
ok wait, dropping a new flagship on a Tuesday with a two-sentence tweet and no thread is such a power move. no launch video, no CEO essay, just numbers go up cost go down.
Keep thinking is a great tagline until my agent loop keeps thinking for 40 minutes on a task that should take 4. reliability numbers or it didn't happen.
Interesting. Would love to see retention on Opus 5 users actually migrating vs staying pinned to the old version for reproducibility. Cheaper doesn't matter if nobody switches.
The screenshot in the tweet has that classic Anthropic beige energy. At this point the muted palette is doing more brand work than the model names are.