Model intelligence / Live profile

Claude Opus 4.8

Track Claude Opus 4.8 from Anthropic across launches, availability, pricing, benchmarks, capabilities, and consequential updates. This page updates automatically as substantive Superpower Daily coverage is published.

Official page

Coverage telemetryActive

Developer / company

Anthropic

Published9

Launches4

Pricing0

Benchmarks1

Latest coverageAnthropic Says Claude Found Fixes Across 10 Alignment Failures, but Tests Remain Narrow

Published coverage9 stories

Last 90 days8 updates

Coverage dates8 days

Profile statusIndexed

Latest coverage

What changed around Claude Opus 4.8.

Browse all models

Sep 15, 2026Tools

Launch Tools

Ryan Greenblatt Launches an AI Hotline That Works Through Restricted Web Access

The service gives agents a route to report information even when they can only fetch web pages. But its fallback method puts messages into URLs, where they can be logged more widely than a standard submission.

3 min

Sep 9, 2026Models

Security risk Models

Anthropic Adds a Fourth Claude Cyber Incident and Recasts the Failures

The newly disclosed case was found during a later transcript review. Anthropic now says a misconfigured test environment and the models’ own behavior both contributed to real-world access.

3 min

Aug 31, 2026Tools

Open release Tools

Vercel Says design.md Cut Known Page Failures 57%, but All Six Pages Still Blocked

The public guidance file is meant to give agents a shared design vocabulary outside a company repository. Its early result is promising, but the small, company-run test still found a shipping-blocking problem in every generated page.

3 min

Aug 31, 2026Models

Open release Models

DeepSeek Releases 305B Vision Weights Under MIT License, but Local Use Needs Serious Hardware

The release gives developers the model files and serving references needed to run or adapt DeepSeek-V4-Flash-Vision-Exp independently. Its 168GB size means that freedom is aimed more at infrastructure-equipped teams than typical individual users.

3 min

Aug 25, 2026Models

Launch Models

DeepSeek V4 Pro Lands on Fireworks With $2.50 CyberGym Solves, but Kimi K3 Scores Higher

The provider’s results frame security-agent selection as a three-way choice among raw solve rate, cost per completed task, and whether a model will reliably execute the required tool workflow.

3 min

Aug 22, 2026Models

Benchmark Models

Stanford Won Databricks’ Agent Cup, but 18.8% of Questions Stumped Every Team

The live contest tested whether agents developed on one benchmark could handle an unfamiliar Treasury archive. Stanford led the field, but the unsolved questions show the limits of reliable document reasoning.

2 min

Aug 22, 2026Models

Launch Models

DeepSeek Gives V4-Flash a Separate Vision API With a 384-Token Image Cap

The experimental model gives developers several familiar ways to submit images and a fixed upper bound on image-token use. It leaves the harder questions—visual accuracy, latency and reliability—for users to test.

3 min

Jun 11, 2026Daily issue