What Is Claude, and What Did Anthropic…

This article summarizes company announcements and press reports as of October 8, 2026. Claude is made by Anthropic, and some items below, such as the IPO and Pentagon reports, are based on headlines and early coverage that may change. Check Anthropic's own pages for current details.

Key Takeaways

Claude is Anthropic's AI assistant, available on the web, mobile, and desktop, and through developer tools including Claude Code, Claude Cowork, Claude in Chrome, Excel, and PowerPoint, and the new Slack-based Claude Tag.
Anthropic's newest releases are Claude Opus 5.5 (October 6), which Anthropic says performs at the level of Claude Fable 5.1 on most work while costing 40% less to run than Opus 5, and Claude Haiku 5.5 (October 7), described as its fastest, cheapest, and most capable small model yet.
The current lineup is Claude Haiku 5.5, Claude Sonnet 5.5, Claude Opus 5.5, and Claude Fable 5.1, with Claude Mythos 5.1, which shares an underlying model with Fable 5.1, available to a limited group of verified organizations.
On October 6, Anthropic expanded its Cyber Verification Program into three tiers, folding in Project Glasswing, whose partners reportedly found more than 129,000 verified software flaws between April and July 2026.
This week also brought scrutiny: a disputed "Claude-led" CRISPR-like discovery, an unconfirmed report about the Pentagon, and a brief outage.

What Claude is, in plain terms

Claude is a family of AI models and the assistant built on them. You can chat with it for writing, research, analysis, and coding help, or use it as an agent that carries out multi-step tasks. For businesses and developers, Anthropic sells access through plans for individuals and teams and through an API that is also offered on Amazon Bedrock, Google Cloud, and Microsoft Foundry. The company says the model is trained with its "constitutional AI" approach, aimed at improving ethical and legal compliance.

Over 2026, the product surface has widened well beyond a chat window. Claude Code handles software work from the command line and desktop, Claude Cowork brings similar agentic help to non-developers, and Claude Design, released in April, generates prototypes, slides, and marketing materials from natural-language prompts. Claude Code Security, introduced in February, reviews codebases for vulnerabilities.

The October model launches

Two launches landed within a day of each other. Claude Opus 5.5 arrived on October 6, with Anthropic saying it matches Claude Fable 5.1 on most work at 40% lower running cost than Opus 5. Claude Haiku 5.5 followed on October 7 as the company's lowest-cost small model, aimed at high-volume, latency-sensitive work. An AI news aggregator reports pricing of $0.10 per million input tokens and $0.50 per million output tokens for prompts up to 100,000 tokens, roughly 75% below Haiku 4.5.

Developers should note a practical warning in Anthropic's release notes: code written for Claude Haiku 4.5 can break on Haiku 5.5, and the same text can count as more tokens. Anthropic also cut the price of prompt cache reads on Claude Sonnet 5.5 from $0.20 to $0.10 per million tokens on October 7, and said Claude Max and Team plans now include monthly API credits.

New tools and programs this month

Several product launches clustered in late September and early October. Claude Tag lets anyone tag @Claude into Slack conversations, a shared, multi-person way of delegating work. Anthropic launched Claude Marketplace, reported to include more than 2,000 plugins, connectors, and partner agents, and released Claude Code Mods, TypeScript functions that can rewrite prompts, tool calls, and parts of the interface. Claude for Government reached general availability on September 30, adding the Claude Code command-line tool and early access to Microsoft 365.

Anthropic is also putting money behind adoption. It committed $100 million to its Claude Frontier Academy, aiming to train 10,000 enterprise AI engineers by 2027. On October 6 it expanded Claude for Startups, offering qualifying companies a free year of Claude Team with up to five premium seats and $1,000 in API credits, for companies founded in the last five years or funded in the last two.

Cybersecurity and access controls

Alongside the models, Anthropic restructured access to its most capable systems. The expanded Cyber Verification Program, announced October 6, consolidates Project Glasswing into three tiers: Defense Access for incident response and vulnerability validation, Red Team Access for authorized penetration-testing organizations, and Specialized Access for safety-critical systems such as flight operating systems and power grids. Qualifying security professionals receive reduced-blocking classifiers and access to Claude Mythos. Bloomberg reported that verified organizations can use the most capable models for high-risk offensive testing of protections around critical infrastructure, in collaboration with the US government.

Access to the newest frontier models has also been interrupted before. Claude Fable 5 and Mythos 5 were first released on June 9, 2026. Anthropic suspended access on June 12 to comply with US Department of Commerce export controls, the Department lifted those controls on June 30, and Anthropic restored access on July 1.

The scrutiny around the headlines

Not all of the week's news was a launch. CNN reported that Anthropic's announcement of a Claude-led discovery of a novel enzyme system with CRISPR-like repeats is drawing controversy. Feng Zhang of MIT and the Broad Institute called the work genuinely intriguing in Anthropic's release, but a University of Copenhagen doctoral researcher, Mario Rodríguez Mestre, says his unpublished research describes the same viral signatures and that he had shared drafts on his private Claude account. He said he is not claiming Anthropic deliberately took his work. The questions about provenance are still open.

Other reports remain thinly sourced. A headline compiled by AI Weekly, citing BBC sources, says the Pentagon has stopped using Claude, and The Information reported a gray market of Chinese resellers routing Claude accounts through proxies. Treat both as unconfirmed until Anthropic or the parties involved say more. Anthropic also published a 2026 usage policy update on October 8, and Claude, Claude Code, and the API had a 40-minute partial outage on September 29. Bloomberg headlines this week also report that Anthropic is targeting an IPO before Thanksgiving, though the sources reviewed did not include company-confirmed details.

What this means if you're choosing or following Claude

If you build on the API, test your prompts before moving to Haiku 5.5. Anthropic's own notes say older code can break and token counts can change, so a low per-token price does not automatically mean a lower bill.
If you want strong results at lower cost, compare Opus 5.5 with Fable 5.1 on your own tasks. Anthropic says they perform similarly on most work, which is worth checking against your own workloads.
If you work in security, look at the new Cyber Verification Program tiers. Access to the strongest models is gated by verification rather than being open to everyone.
Either way, read claims about discoveries and government use carefully. Several of this week's biggest stories involve early reporting or disputes that are still developing.

Frequently Asked Questions

What is Claude? Claude is Anthropic's AI assistant and model family. You can use it for writing, analysis, research, and coding, or as an agent that completes multi-step tasks, through apps, the API, and tools like Claude Code and Claude Cowork.

What is Claude's newest model? Claude Haiku 5.5 launched on October 7, 2026, and Claude Opus 5.5 on October 6. The current lineup also includes Claude Sonnet 5.5 and Claude Fable 5.1, with Claude Mythos 5.1 limited to verified organizations.

What is Claude Tag? Claude Tag is a Slack-based interface where people can tag @Claude into a conversation and delegate tasks to it alongside colleagues.

Is Anthropic going public? Bloomberg headlines this week say Anthropic is targeting an IPO before Thanksgiving. The sources reviewed did not include confirmed details from the company, so check for official announcements.

Sources & References

Anthropic, Newsroom (October 2026 announcements)
Claude Platform Docs, release notes (October 7–8, 2026)
TechCrunch, "Anthropic is giving startups a free year of Claude Team and $1,000 in credits"
Bloomberg, "Anthropic Grants US Cyber Firms Access to Latest Claude AI Models"
CNN, "Anthropic's announcement of a 'Claude-led' CRISPR-like discovery is sparking controversy"
AI Weekly, "AI News Today, October 8: Top Stories"
Anthropic, "Fable and Mythos access" statement

What Is Claude, and What Did Anthropic Just Launch in October 2026?