AI incidents: Claude and Runway recoveries, Suno download concerns — 25 October 2026

Publisher disclosure: MUVO offers competing music, remix, and video tools. This report summarizes public sources and provider documentation; it is not an independent product test or a claim that MUVO is free of similar issues. Read our reporting method.

Incident summary and status at publication

Anthropic has resolved two distinct Claude error incidents on 25 October, and Runway says its Magnific upscalers recovered on 24 October. Separate customer reports concern Suno’s restrictions on downloading older songs and ElevenLabs billing after reported cancellation. The customer accounts do not establish service-wide outages.

This edition covers 22 October at 12:29 through 25 October at 11:56 Europe/Istanbul (UTC+3). The Runway recovery continues yesterday’s coverage; it is not a new outage. All event times below are Istanbul time, with original UTC timestamps retained in our evidence ledger.

Claude: two separate error incidents are resolved

Anthropic gives actual impact windows of 30 minutes and 21 minutes, rather than requiring an estimate from the announcement times. Both records associate claude.ai, the Claude API, Claude Code and Claude Cowork. Their scope is errors affecting the named models; the component labels do not establish that every request across those products failed.

Sonnet 5.5 and other models: 09:05–09:35

Anthropic’s first acknowledgement at 09:14 named Mythos 5.1, Fable 5.1, Opus 5.5, Sonnet 5.5 and Haiku 5.5. An identification update at 09:27 focused on Sonnet 5.5. At 09:35, the provider said a fix was deployed and monitoring had begun. Its later 10:53 resolution notice states that impact lasted from 09:05 to 09:35. Original incident and updates.

Fable 5.1: 11:08–11:29

A separate Fable 5.1 incident was acknowledged and identified at 11:17. A fix entered monitoring at 11:30. Anthropic’s 11:49 resolution notice retrospectively gives the impact window as 11:08–11:29. Monitoring began after that stated impact end; neither the first notice nor the final resolution timestamp is the full customer-impact boundary. Original Fable incident.

Anthropic says it identified the causes but does not disclose technical details, affected regions or a user workaround in these notices. No shared cause is established. Provider resolution does not mean that every failed request automatically completed or every individual account recovered.

Runway: Magnific upscaling recovered after yesterday’s cutoff

Runway reported Magnific upscaling working again at 15:21 on 24 October and closed the incident at 16:29. The provider said it had applied a fix and was monitoring stability before resolution. These updates came after yesterday’s 12:29 research cutoff. Original incident, timestamped public record.

The initial notice was at 12:15, roughly 3 hours 6 minutes before the recovery statement. That measures the public notice-to-recovery interval, not the complete outage duration: the actual impact start is unknown. The initial notice referred to work with an unnamed provider but disclosed no technical cause. The acknowledged failure concerned Magnific upscaling, not all Runway video generation. No source-supported workaround was supplied.

We have added a dated recovery note to the 24 October article, retaining its original publication date, URL and original-cutoff timeline.

Suno: a customer reports restrictions on downloading older paid creations

A review dated 24 October describes difficulty exporting older songs after a paid subscription ended, and asks how commercial-use rights apply to those songs. This is a customer-reported access concern, not evidence of a newly introduced outage. The account appears on Suno’s original review page; its individual permalink was not separately retrievable.

Suno’s own FAQ corroborates the relevant download restriction: it says limits apply to older music from 3 September, and cancelled accounts follow free-tier download limits. It also says songs remain playable and shareable in the library. This is an existing policy, not a change announced today. Provider FAQ, dated 10 August.

There is a documentation distinction behind the customer’s question. Suno’s September guidance refers to songs downloaded while subscribed, while an older December 2025 help page refers to songs made while subscribed. Current paid-subscription guidance, older distribution guidance. Those pages do not settle the reviewer’s individual entitlement or establish that previously granted rights were revoked. No account-specific provider reply or resolution was verified. We cannot assign an outage duration to a plan restriction.

ElevenLabs: two customers report billing concerns after cancellation

Two separate reviewers dated 23 October say subscriptions appeared to restart after cancellation. One reports a renewed charge; the other describes an attempted charge and difficulty obtaining help. These are customer allegations, without provider confirmation or independently verified payment records. Original review page, individual references one and two.

The individual pages were restricted, but both full accounts and their dates were readable on the original company review page. Different profiles provide matching firsthand descriptions; account state, cancellation channel and charge timing remain unknown. No shared technical cause, fraud finding, prevalence or company-wide payment outage is established.

Counterevidence on the same page includes positive product use dated 23 October and a support-assisted resolution of a different case dated 21 October. Neither resolves these two accounts. No provider reply to the matching complaints or reliable workaround was verified; the incidents’ start and end times are unknown.

Follow-up checks and coverage limits

OpenAI’s status history had no newly verified event beyond those covered in yesterday’s report. An isolated new account of ChatGPT HTTP 429 errors remains an unpublished lead; it does not prove a general outage. Older account-capacity, Work-loading and dictation threads did not establish a new general recovery. OpenAI history.

Google’s London report remains preliminary, without a newly verified final analysis. ElevenLabs’ public history yielded no fresh substantive service incident; previously promised write-ups remained unreadable. These checks are separate from the billing allegations above. Google incident, ElevenLabs history.

Restricted review feeds, incomplete xAI, DeepSeek and Perplexity histories, and unavailable private Microsoft tenant notices limit coverage. Consumer Gemini history was not fully accessible. A healthy current indicator is not evidence that no disruption occurred earlier.

Sources

• Anthropic’s multi-model incident and Fable 5.1 incident.
• Runway’s Magnific incident history.
• Original Suno and ElevenLabs customer accounts, with provider documentation cited beside the relevant claims.

For launches and releases, see today’s product-news roundup. MUVO publishes this reporting and offers competing creative tools. Public reviews are not proof of purchase or independently audited account histories; see our reporting method.

Update timeline

• 25 October 2026, 11:56 Istanbul / 08:56 UTC: Research cutoff. Both new Claude records were resolved. Added Runway’s verified recovery and a dated update to yesterday’s article; separately labeled substantiated customer-report issues without treating them as confirmed outages.