Skip to content
DailyPrompt

your daily dose of claude

01 announcement

Anthropic Releases Claude Opus 4.7, Its Most Capable Publicly Available Model

Anthropic launched Claude Opus 4.7 on April 16, delivering its strongest generally available model to date with significant leaps across coding, vision, instruction-following, and long-running agentic tasks. Standout improvements include a 13% uplift in coding task resolution over Opus 4.6, 21% fewer document reasoning errors, and a vision accuracy jump from 54.5% to 98.5% thanks to native support for images up to 3.75 megapixels. The model is available immediately on Claude.ai and through the API, as well as Amazon, Google, and Microsoft cloud platforms, at the same $5/$25 per million token pricing as Opus 4.6.

Anthropic →
research

Bloomberg: How Anthropic's Own Safety Tests Convinced It to Lock Away Claude Mythos

A Bloomberg feature reconstructs the internal Anthropic debate that led the company to withhold Claude Mythos from public release entirely. Rather than external regulatory pressure, it was the team's own safety evaluations — showing Mythos could autonomously complete end-to-end cyberattacks — that drove the decision to restrict it to vetted partners through Project Glasswing. The piece is one of the first detailed accounts of an AI lab using its own red-team findings to actively constrain, rather than merely caveat, a model it had built.

Bloomberg
C
Claude
@claudeai

Introducing Claude Opus 4.7, our most capable Opus model yet. It handles long-running tasks with more rigor, follows instructions more precisely, and verifies its own outputs before reporting back. You can hand off your hardest work with less supervision.

Apr 16 View on X →
product

$ cat story_03.md

OpenAI Fires Back at Claude Code with a Revamped Codex That Controls Your Desktop

Hours after Anthropic released Opus 4.7, OpenAI announced a major overhaul of its Codex coding agent — adding the ability to run background tasks and take actions directly on a user's desktop. The timing was clearly deliberate: Claude Code has been the dominant AI coding tool in enterprise settings for months, and OpenAI is positioning the new Codex as a direct response. The two products now compete feature-for-feature on agentic capabilities, persistent sessions, and IDE integration.

$ open techcrunch →
community No. 04

Anthropic Acknowledges Claude Effort Changes But Power Users Remain Unconvinced

As Anthropic launched Opus 4.7, the fallout from its quiet adjustments to Claude's default effort levels continued to grow. Despite the company acknowledging it made configuration changes affecting output depth, the explanation has done little to pacify the heavy users who first noticed the degradation. Critics argue the company's admission that it reduced default thinking depth to manage compute costs — without any public announcement — represents a fundamental breach of trust with its most loyal paying customers.

Source: Axios →
A
Alex Albert
@alexalbert__

Some of my favorite things in Opus 4.7: - Very good at async work and following instructions - Effort levels are far more predictable for token control (+ new xhigh level) - No more downscaling of high-res images - Noticeably more taste in UIs, slides, docs

Apr 16 View on X →
05
community

Anthropic Sends Legal Threats to OpenCode Over Subscription Token Access

Anthropic has threatened legal action against OpenCode, a popular third-party Claude Code client, over its plugins that allowed users to access Claude Pro and Max subscription tokens through the unofficial interface. OpenCode promptly released version 1.3.0 with those features removed. The move continues Anthropic's enforcement push against third-party tools that use cheaper subscription pricing, following earlier clashes with OpenClaw and the suspension of its creator's account.

Hacker News →
T
Thariq
@trq212

We've heard your feedback and we're working on making it easier to follow everything that's happening with Claude Code. First, we're introducing @ClaudeDevs, the official channel to follow for all updates on Claude Code and the Claude platform.

Apr 17 View on X →
product

Anthropic Launches @ClaudeDevs, a Dedicated Developer Channel for Claude Code Updates

Anthropic launched @ClaudeDevs, a new official X account specifically for developers building with Claude, offering a direct line to changelogs, API release notes, community updates, and technical deep dives. The launch came in direct response to developer feedback that Claude Code news was too scattered across multiple accounts. With Anthropic's engineering team posting directly from the account, it signals a more deliberate effort to treat developers as a distinct audience rather than routing everything through the general @AnthropicAI channel.

Read on X / ClaudeDevs →
F
Felix Rieseberg
@felixrieseberg

In the desktop app, Claude Cowork and Code now have a little Bluetooth API for makers & developers, allowing you to build hardware devices that interact with Claude. You can find the API reference and example code here: https://github.com/anthropics/claude-desktop-buddy

Apr 17 View on X →