Anthropic launched Claude on March 14, 2023. Since then, the AI assistant has grown from a single research-preview model into a full family of AI tools covering chat, coding, agents, vision, and enterprise workflows.
This article covers every major Claude model and update, with dates, data, and assessments of what each release actually changed.
Claude Fable 5 and Mythos 5
Date: June 9, 2026
New Model Tier
Key Features
Claude Fable 5 is Anthropic's first publicly available Mythos-class model, it scores 80.3% on SWE-Bench Pro (11 points ahead of the next-best frontier model), leads GDP.pdf's vision benchmark at 29.8%, and delivers state-of-the-art performance on long, complex agentic tasks.
Other Key Features:
- Fable 5 is Anthropic's first public Mythos-class model, sitting above Opus 4.8 and posting state-of-the-art scores on nearly every benchmark tested, with the biggest lead on long, complex coding tasks.
- Two variants exist: Fable 5 (public, with safety classifiers active) and Mythos 5 (restricted to vetted cybersecurity and biology partners), sharing the same underlying weights.
- API pricing is $10/1M input and $50/1M output tokens, a step up from Opus 4.8, reflecting the added capability tier.
Claude Design
Date: April 17, 2026
New Product
Key Features
Claude Design is a collaborative visual creation tool released by Anthropic Labs on April 17, 2026, that lets users generate designs, prototypes, slides, and marketing materials through natural-language prompts. It is powered by Claude Opus 4.7.
Other Key Features:
Brand-Aware Generation: Claude Design can pull brand systems from a user's codebase or design files.
Flexible Iteration: Users can iterate on outputs through inline comments, direct edits, and adjustment sliders.
Broad Export Options: Results export to Canva, PDF, PPTX, HTML, or folders, with a handoff path to Claude Code for development.
Claude Opus 4.8
Date: May 28, 2026
New Model
Key Features
Claude Opus 4.8 shifts the emphasis toward honesty and reliability. Anthropic reports it is roughly four times less likely than Opus 4.7 to let flaws in code it has written pass unremarked.
Other Key Features:
- Debugger: One persistent frustration with earlier Opus models was confident code generation that introduced bugs without flagging them. That was reduced fourfold.
- Strong Agent Performance: Opus 4.8 scored highly on agentic task handling and professional work benchmarks.
- Long Sessions: It became the recommended model for long-running collaborative workflows, particularly where consistency across many steps matters more than raw speed.
Claude Opus 4.6
Date: February 5, 2026
New Model
Key Features
Claude Opus 4.6 launched as the most capable model in the lineup. It boasted a one-million-token context window in beta and improved multi-step reasoning that breaks tasks into subtasks and runs them in parallel.
Other Key Features:
- Legal Reasoning Breakthrough: Opus 4.6 scored 90.2% on BigLaw Bench for legal reasoning. Legal reasoning is one of the hardest domains for AI, given ambiguous language, nested conditions, and high stakes for errors.
- Finance: It posted leading scores on finance tasks covering due diligence and market intelligence.
- Native Multi-Agent Collaboration: Anthropic introduced Agent Teams alongside this model, enabling multiple Claude agents to coordinate on a single task for the first time.
Claude Cowork
Date: January 2026
New Tool
Key Features
Claude Cowork is a tool similar to Claude Code but with a graphical user interface for non-technical users. It was released in January 2026 as a research preview, available as a desktop application for Mac and Windows and through a web interface.
Other Key Features:
- Local File Access: Cowork provides Claude with access to a sandboxed shell and to user-selected folders on the local file system, allowing the model to read, write, and edit files, execute code, and chain multi-step tasks within a single conversation.
- Enterprise Connectors: Anthropic added connectors in February 2026 for services such as Google Drive, Gmail, DocuSign, and FactSet. It also set up customizable plugins for domains, including financial analysis, engineering, and human resources.
Claude Opus 4.5
Date: November 2025
Model Update
Key Features
Claude Opus 4.5 delivered a 67% price cut and 76% fewer output tokens, making premium intelligence affordable for everyday use.
Other Key Features:
Increased affordability: This was the release that removed the cost barrier separating startups and individual developers from Opus-tier intelligence.
New Restrictions: The 76% reduction in output tokens occasionally produced responses that were overly terse and required follow-up prompts to get full answers.
Claude 4
Date: May 2025
New Model Generation
Key Features
The Claude 4 generation brought professional coding ability that made Claude Code a daily driver for serious development work.
Other Key Features:
1-Million-Token Context (Beta): Gave developers room to work across entire large codebases and lengthy documents in a single session.
Effort: Let developers trade speed against capability, a level of fine-grained cost and latency control competitors weren't offering at the time.
Memory Tool: An explicit memory tool gave Claude continuity across sessions instead of starting from zero each time.
Claude Code
Date: February 24, 2025
New Agentic Tool
Key Features
Claude Code came out on February 24, 2025, as a limited research preview with Claude 3.7 Sonnet. It was a coding agent that could work directly with repositories, run commands, edit files, and handle implementation tasks from the command line.
Other Key Features:
Real Time Savings: In early testing, Claude Code completed tasks in a single pass that would normally take over 45 minutes of manual work.
Increasing Capabilities: Later updates added remote MCP server support, web access, subagents, routines, and dynamic workflows.
Claude 3.7 Sonnet
Date: February 2025
New Model + Reasoning Capability
Key Features
Extended thinking with Claude 3.7 Sonnet introduced hybrid reasoning that let Claude pause and think step-by-step before responding. There was improvement in the quality of complex math, science, and multi-step code problems.
Claude 3.5 Sonnet
Date: June 20, 2024
New Model
Key Features
Claude 3.5 Sonnet delivered Opus-level performance at Sonnet pricing. In my view, this was the single most significant Claude release to date; it changed who could afford to build with Claude.
Other Key Features:
Computer Use (October 2024 Update): Anthropic released an updated version with Computer Use in public beta. Replit used the capability to develop a feature that evaluates apps as they're being built for its Replit Agent product.
Automation: The Browser Company, using the model for automating web-based workflows, noted Claude 3.5 Sonnet outperformed every model they had tested before.
Claude 3 Family
Date: March 4, 2024
New Model Family
Key Features
Claude 3 was Anthropic's most important product moment up to that point. Releasing three models simultaneously, Haiku, Sonnet, and Opus, let Anthropic compete on price at the low end while maintaining a credible frontier claim at the top.
Other Key Features;
- Three-Tier Structure: Since Claude 3, each generation has been released in three sizes, from least to most capable: Haiku, Sonnet, and Opus.
- Multimodal Processing: Multimodal capability arrived at every price point for the first time, not just the flagship model.
- Generally Available Tool Use: Combined with vision, this gave enterprise developers a complete stack to build on for the first time.
Claude Pro
Date: September 7, 2023
Paid Plan Launch
Key Features
Anthropic introduced a paid plan for the Claude.ai experience, giving subscribers priority access, faster responses, and early entry to new features.
Claude 2
Date: July 11, 2023
New Agent
Key Features
Claude 2, released on July 11, 2023, became the first Anthropic model available to the general public.
Other Key Features:
200K Context Window via Claude 2.1: The follow-up release pushed the context window to 200K tokens, far beyond what competitors offered at the time.
New Enterprise Use Cases Unlocked: That context length made it possible to process full legal contracts, entire codebases, and lengthy financial filings in a single pass.
Lower Hallucination Rate: Claude 2.1 also reduced hallucinations and added beta tool use and system prompts.
100K Context Window
Date: May 11, 2023
Update
Key Features
Claude was the first foundation model from a frontier lab to reach a 100K-token context window.
Other Key Features:
New Workflows: At the time, most models topped out at 4K or 8K tokens, so this jump unlocked entirely new workflows.
Enterprise Clientele: Legal teams, researchers, and developers working with large codebases adopted Claude quickly, since no alternative existed at that scale.
Claude 1
Date: March 14, 2023
First Public Model
Key Features
The first publicly available Claude model arrived on March 14, 2023, when Anthropic launched the Anthropic API with Claude and Claude Instant. Access was limited to approved users and API partners.
Other Key Features:
Constitutional AI: Claude was designed around helpful, harmless, and honest principles from day one.
Coding and Writing: Claude handled conversation, summarization, coding, Q&A, and creative writing competently at launch.
Anthropic's Founding and Claude
Date: January 2021
New Model
Anthropic was founded in 2021 by Dario Amodei, Daniela Amodei, and a group of former OpenAI researchers to build AI systems that were safe, steerable, and reliable.
