How to read this timeline
The archive treats a model launch, a feature release, a deprecation notice, and a retirement as different events. A launch page tells you when Anthropic presented a model; the deprecation ledger tells you when a model stopped being safe to target in production; the migration map records the practical upgrade pressure that followed.
This page is intentionally historical. For week-to-week news, use Claude Weekly; for Claude Code engineering practice, use Claude Ships Code.
Chronological ledger
| Date | Event | Why it mattered | Source |
|---|---|---|---|
| Claude and Claude Instant launch | Anthropic introduces Claude more broadly after a closed alpha with partners, offering Claude and Claude Instant through chat and API surfaces. | Introducing Claude | |
| Claude 2 reaches public beta | Claude 2 arrives with API access, claude.ai beta access in the US and UK, and a 100K-token input window. | Claude 2 | |
| Claude 2.1 expands context to 200K | Claude 2.1 adds a 200K-token context window, system prompts, beta tool use, and an updated pricing schedule. | Introducing Claude 2.1 | |
| Claude 3 family launches | Claude 3 debuts as Haiku, Sonnet, and Opus, making the family names a persistent structure for later releases. | Introducing the next generation of Claude | |
| Claude 3.5 Sonnet launches | Anthropic launches the first Claude 3.5 model at the Sonnet tier with a 200K context window and $3/$15 per MTok API pricing. | Claude 3.5 Sonnet | |
| Computer use and Claude 3.5 Haiku arrive | Anthropic announces an upgraded Sonnet 3.5, Claude 3.5 Haiku, and public beta computer use. | Introducing computer use, a new Claude 3.5 Sonnet, and Claude 3.5 Haiku | |
| Claude 3.7 Sonnet and Claude Code | Claude 3.7 Sonnet introduces hybrid reasoning while Claude Code appears as a limited research preview. | Claude 3.7 Sonnet and Claude Code | |
| Claude 4 launches | Claude Opus 4 and Claude Sonnet 4 launch with stronger coding, agent, and reasoning capabilities; Claude Code becomes generally available. | Introducing Claude 4 | |
| Claude Opus 4.1 | Opus 4.1 upgrades Opus 4 for agentic tasks, real-world coding, and reasoning while keeping Opus 4 pricing. | Claude Opus 4.1 | |
| Claude Sonnet 4.5 | Sonnet 4.5 launches with major coding, agent, computer-use, and reasoning gains and retains Sonnet 4 pricing. | Introducing Claude Sonnet 4.5 | |
| Claude Haiku 4.5 | Haiku 4.5 launches as the speed and cost-efficiency member of the modern family. | Introducing Claude Haiku 4.5 | |
| Claude Opus 4.5 | Opus 4.5 launches on apps, API, and major cloud platforms with $5/$25 per MTok pricing. | Introducing Claude Opus 4.5 | |
| Claude Opus 4.6 | Opus 4.6 launches with a 1M-token context-window beta for the Opus class. | Claude Opus 4.6 | |
| Claude Sonnet 4.6 | Sonnet 4.6 becomes the default model for Free and Pro users and adds a 1M-token context-window beta. | Introducing Claude Sonnet 4.6 | |
| Claude Opus 4.7 | Opus 4.7 launches with new cybersecurity safeguards and the same $5/$25 per MTok pricing as Opus 4.6. | Introducing Claude Opus 4.7 | |
| Claude Opus 4.8 | Opus 4.8 launches with effort controls, dynamic workflows, and cheaper fast mode than prior Opus fast mode. | Introducing Claude Opus 4.8 | |
| Claude Fable 5 and Mythos 5 | Anthropic launches Fable 5 for general use and Mythos 5 for limited approved cybersecurity access. | Claude Fable 5 and Claude Mythos 5 | |
| Claude Sonnet 5 | Sonnet 5 launches as the newest Sonnet model with introductory $2/$10 per MTok pricing through August 31, 2026. | Introducing Claude Sonnet 5 |
The shape of the model line
The first visible era was Claude plus Claude Instant. The second was Claude 2 and 2.1, where context-window size and enterprise-readiness became the headline. The third introduced persistent family names: Haiku for speed and cost, Sonnet for the middle tier, and Opus for the highest-capability tier.
The 2025 and 2026 eras shifted the emphasis toward agentic work. Claude 3.7 Sonnet introduced hybrid reasoning and Claude Code. Claude 4 made Opus and Sonnet the main agent and coding story. The 4.5, 4.6, 4.7, and 4.8 releases then tightened pricing, context, tool use, computer use, and safety posture. The 5-era pages introduced Fable, Mythos, and Sonnet 5 as new names in the source record.
Known gaps and limits
This page does not treat leaked prompts, scraped app behavior, or third-party benchmark mirrors as canonical history. If an event cannot be tied back to an Anthropic-owned source, it belongs in a notes queue, not in the archive table.
Partner platform availability can differ. The model-deprecation docs explicitly distinguish Anthropic-operated platforms from partner-operated platforms such as Amazon Bedrock and Google Cloud, so retirement dates here should not be reused for those platforms without checking their own tables.
Questions this page answers
When did Claude first launch?
Anthropic publicly introduced Claude on March 14, 2023, after a closed alpha with partners.
What was the first Claude 3 family release date?
Anthropic announced the Claude 3 family on March 4, 2024.
What is the latest Claude timeline event in this archive?
As verified on July 6, 2026, the latest Sonnet event is Claude Sonnet 5 on June 30, 2026; Fable/Mythos 5 access restoration is recorded as July 1, 2026 in Anthropic launch/update pages.