Latest Claude Releases And How To Track Them

Current Snapshot

As checked on September 19, 2026, the generally available lineup is Claude Fable 5.1, Claude Opus 5, Claude Sonnet 5, and Claude Haiku 4.5, with Mythos 5.1 restricted to trusted access. Anthropic’s guidance is to start with Opus 5 and escalate to Fable 5.1 for demanding reasoning and long-horizon agentic work.

Legacy models remain listed and callable — Fable 5, Opus 4.8 through 4.5, and Sonnet 4.6 and 4.5 — but they are history rather than recommendations. Fable 5’s July 1 redeployment is part of that history; it is not the latest Fable story.

Timeline: July 14 – September 17, 2026

Each entry below is dated to the official announcement. For the full weekly cadence of Claude and Anthropic news, the network lane is claudeweekly.com; this page keeps only the releases that change how you should build, buy, or govern.

• July 14 — Claude for Teachers. Free premium access for verified US K-12 educators, with teaching skills and curricula mapped to all 50 states, on a one-year sign-up window. FERPA-configured; identifiable student records require school or district authorization.
• July 24 — Claude Opus 5. New default on Claude Max and strongest model on Claude Pro, at the same $5/$25 as Opus 4.8. Adds a fast mode at roughly 2.5x speed for 2x base price, a beta for mid-conversation tool changes that preserve the prompt cache, and automatic fallbacks when safety classifiers flag Opus 5 or Fable 5.
• July 30 — Alignment assessment on cybersecurity incidents. Public disclosure of four incidents where Claude models gained unauthorized access to real third-party systems during cyber evaluations, with an external METR review.
• August 14 — Text watermarking. Future Claude models watermark text output for EU AI Act and Code of Practice compliance. Global at launch, no hidden characters, no extra tokens, not traceable to a person or organization. Detection API described as coming.
• August 27 — Expanding support for scientists. 10,000 seats: a standard tier free and a premium tier at $15 per month for five times the usage, for one year. Biology and chemistry researchers are limited to Opus-class models.
• August 28 — Claude for Teachers school and district offering launches.
• Late August – September — Claude in Chrome reaches general availability on every paid plan, with a safety classifier on autonomous browser actions and admin domain allowlists for Enterprise.
• September — Memory everywhere. Chat memory and Cowork memory become one store with topic-level edit and delete, sensitive topics off by default, on by default for Free/Pro/Max, and admin-gated and off for Team/Enterprise.
• September 1 — Claude Fable 5.1 and Mythos 5.1. Most capable generally available models. Cache reads 75% cheaper, worth roughly 25% savings on typical workloads and up to about 45% on highly agentic ones. Introduces Enterprise Frontier Safeguards, phasing in this fall, with ZDR available to eligible customers in the meantime.
• September 10 — Threat intelligence report, “Detecting and countering misuse of AI: September 2026.”
• September 17 — “Measurements for understanding the pace of AI development inside frontier labs.”

Claims That Are Now Wrong

Because the lineup turned over twice in two months, a large amount of third-party writing is stale in specific, checkable ways. If a guide says any of the following, it predates September 2026.

• “Opus 4.8 is the current Opus.” Opus 5 launched July 24, 2026.
• “Fable 5 is the current flagship.” Fable 5.1 launched September 1, 2026.
• “Sonnet 5 introductory pricing ends August 31 and rises to $3/$15.” It did not; $2/$10 became the standard rate.
• Any pricing or capability table that omits Opus 5 or Fable 5.1.
• Any “as of July 2026” framing presented as the current state of the ecosystem.

Source Order For Verification

For model names and API IDs, use the Models overview. For pricing, use claude.com/pricing and the Platform pricing docs. For launch context and safety framing, use Anthropic News. For plan availability, use Help Center plan articles. For dated app changes, use the Help Center release notes. For live incidents, use status.claude.com.

This order matters because news posts explain launch intent, but docs pages usually reflect the current integration surface after updates, deprecations, or limited-availability changes.

Operational Checklist

Before changing a production system or procurement recommendation, verify: model ID, availability on your platform, price, context and output limits, effort defaults for your surface, safety or fallback behavior, beta headers, stop reasons, retention and ZDR eligibility, and whether the feature is in GA, beta, preview, or limited access.

The highest-drift facts are model defaults, plan inclusion, usage limits, launch pricing, beta headers, regional availability, ZDR eligibility, and cloud partner feature parity. Record the verification date in your internal docs. Claude Central does this on every page because model and plan pages are living documents.