AI provider / Anthropic
Anthropic
A provider frequently associated with careful reasoning, document-heavy work, long context handling and enterprise safety positioning.
Anthropic is the AI company built around reliability and steerability, and the home of the Claude model family. Founded on Constitutional AI, it has grown from the first Claude in 2023 into one of the most trusted names for analysis, long-context work and enterprise-grade reasoning.
Its strategy is a steady climb up a single, well-shaped ladder: the Opus, Sonnet and Haiku tiers, refined release after release from Claude 3 to the 4-series and now the Claude 5 family. Each step pushes coding, agentic work and long-horizon reliability, while keeping safety positioning at the core of the brand.
Model catalog
Every release in this provider’s model line.
Each row is a distinct model release — its date, its type and the traits it is known for. Open a source to dig into any of them.
| Model | Released | Type | Key traits | Source |
|---|---|---|---|---|
| Claude Sonnet 5 | Jun 2026 | Model release |
| Anthropic |
| Claude Fable 5 | Jun 2026 | New family |
| Anthropic |
| Claude Opus 4.8 | May 2026 | Flagship |
| Anthropic |
| Claude Sonnet 4.6 | Feb 2026 | Version update |
| Anthropic |
| Claude Opus 4.5 | Nov 2025 | Flagship |
| Anthropic |
| Claude Haiku 4.5 | Oct 2025 | Small tier |
| Anthropic |
| Claude Sonnet 4.5 | Sep 2025 | Version update |
| Anthropic |
| Claude Opus 4 & Sonnet 4 | May 2025 | Flagship |
| Anthropic |
| Claude 3.7 Sonnet | Feb 2025 | Hybrid reasoning |
| Anthropic |
| Claude 3.5 Sonnet & Haiku (upgrade) | Oct 2024 | Version update |
| Anthropic |
| Claude 3.5 Sonnet | Jun 2024 | Flagship |
| Anthropic |
| Claude 3 (Opus/Sonnet/Haiku) | Mar 2024 | Model family |
| Anthropic |
| Claude 2.1 | Nov 2023 | Version update |
| Anthropic |
| Claude 2 | Jul 2023 | Version update |
| Anthropic |
| Claude & Claude Instant | Mar 2023 | Debut |
| Anthropic |
Timeline
A practical timeline of model releases and company milestones.
This page tracks the moments that matter most: major model releases, version jumps and company moves that changed how this provider shows up in the AI market.
Claude Sonnet 5 brings frontier agentic work into Anthropic’s default tier.
Anthropic positions Sonnet 5 close to Opus 4.8 on coding, tools and professional work, while making it the default model across plans at a lower price point.
Claude Fable 5 opens the Claude 5 family and the Mythos tier.
A new class above Opus arrives in two forms: generally available Fable 5 with added safeguards, and trusted-access Mythos 5.
Claude Opus 4.8 pushes the 4-series to its ceiling.
The last and strongest of the Opus 4 line, tuned for deep coding sessions and large-codebase reasoning ahead of the 5 era.
Claude Sonnet 4.6 closes the gap to Opus on agentic work.
The default tier inherits most of Opus 4.5’s planning ability, continuing the pattern of mid-tier releases eating the flagship.
Claude Opus 4.5 retakes the frontier with a two-thirds price cut.
Opus stops being the luxury option: frontier reasoning at drastically lower cost resets expectations for the whole top tier.
Claude Haiku 4.5 brings near-frontier coding to the small tier.
Matching Sonnet 4 at a third of the cost, the small model becomes the workhorse for high-volume and sub-agent workloads.
Claude Sonnet 4.5 leads on coding while Opus 4.1 holds the high end.
After August’s Opus 4.1, the mid-tier overtakes it for everyday work — thirty-hour autonomous runs at Sonnet pricing.
Claude Opus 4 and Sonnet 4 target long-horizon agent work.
Hours-long autonomous coding sessions become the headline capability, with Opus 4 sold as the best coding model in the world.
Claude 3.7 Sonnet debuts hybrid reasoning and Claude Code.
One model spans instant replies and extended thinking, while a terminal agent quietly starts Anthropic’s agentic-coding franchise.
Upgraded 3.5 Sonnet and 3.5 Haiku ship with computer use.
Claude becomes the first frontier model to drive a desktop — clicking, typing and browsing — in public beta.
Claude 3.5 Sonnet outruns Opus at mid-tier price, with Artifacts.
The release that made Claude a developer favorite: frontier coding skill plus a workspace for running what it writes.
Claude 3 introduces the Opus, Sonnet and Haiku tiers.
The three-tier naming and vision support arrive together, with Opus briefly taking the overall benchmark crown from GPT-4.
Claude 2.1 doubles context to 200K and halves hallucinations.
An enterprise-flavored update with tool use in beta, sharpening Claude’s positioning around reliability over flash.
Claude 2 brings 100K context and a public chat product.
Long documents become Claude’s signature: the 100K window makes contract and report analysis its clearest differentiator.
Claude and Claude Instant open Anthropic’s model line.
Trained with Constitutional AI, the first Claude pair positions helpfulness and harmlessness as a product feature, not a disclaimer.