Gate News message, April 23 — Anthropic’s engineering team confirmed that the Claude Code quality degradation reported by users over the past month stemmed from three independent product-layer changes, not from API or underlying model issues. The three problems were fixed on April 7, April 10, and April 20 respectively, with the final version now at v2.1.116.
The first change occurred on March 4, when the team reduced the default reasoning effort level for Claude Code from “high” to “medium” to address occasional extreme latency spikes in Opus 4.6 under high reasoning intensity. After widespread user complaints about reduced performance, the team reverted the change on April 7. The current default is now “xhigh” for Opus 4.7 and “high” for other models.
The second issue was a bug introduced on March 26. The system was designed to clear old reasoning records after conversation inactivity exceeded one hour to reduce session recovery costs. However, a flaw in implementation caused the clearing to execute repeatedly on every subsequent turn rather than once, causing the model to progressively lose prior reasoning context. This manifested as increasing forgetfulness, repeated operations, and abnormal tool invocations. The bug also resulted in cache misses on every request, accelerating user quota consumption. Two unrelated internal experiments masked the reproduction conditions, extending the debugging process to over a week. After fixing on April 10, the team reviewed problematic code using Opus 4.7 and found that Opus 4.7 could identify the bug while Opus 4.6 could not.
The third change launched on April 16 alongside Opus 4.7. The team added instructions to the system prompt to reduce redundant output. Internal testing over several weeks showed no regression, but post-launch interaction with other prompts degraded coding quality. Extended evaluation revealed a 3% performance drop in both Opus 4.6 and 4.7, leading to a rollback on April 20.
These three changes affected different user groups at different times, and their combined effect created widespread and inconsistent quality decline, complicating diagnosis. Anthropic stated it will now require more internal employees to use the same public build version as users, run full model evaluation suites for every system prompt modification, and implement staged rollout periods. As compensation, Anthropic has reset usage quotas for all subscription users.
Disclaimer: The information on this page may come from third parties and does not represent the views or opinions of Gate. The content displayed on this page is for reference only and does not constitute any financial, investment, or legal advice. Gate does not guarantee the accuracy or completeness of the information and shall not be liable for any losses arising from the use of this information. Virtual asset investments carry high risks and are subject to significant price volatility. You may lose all of your invested principal. Please fully understand the relevant risks and make prudent decisions based on your own financial situation and risk tolerance. For details, please refer to
Disclaimer.
Related Articles
Cognition AI Raises Funding at $25B Valuation in Early-Stage Negotiations
Gate News message, April 24 — Cognition AI, an AI coding startup, is in early-stage negotiations for a new funding round that would more than double its valuation to $25 billion, according to sources familiar with the matter. The company aims to raise hundreds of millions of dollars or more as
GateNews42m ago
NEC Corporation becomes Anthropic’s first global partner in Japan
NEC announced it will become Anthropic’s first global partner in Japan. The two companies will develop secure, industry-knowledgeable AI solutions for highly regulated industries such as finance, manufacturing, and local governments, and integrate the Claude series into NEC BluStellar. They will focus on data-driven management and transforming customer experience, while also introducing Claude Cowork and SOC integration to enhance cybersecurity protection. To validate effectiveness, NEC launched the Zero Customer Program to conduct comprehensive internal testing of AI agents, and plans to promote Claude deployments globally—building Japan’s largest-scale AI-native engineering CoE.
ChainNewsAbmedia3h ago
Vercel Security Breach Expands to Hundreds of Users; AI Developers at Higher Risk
Gate News message, April 23 — Vercel disclosed on April 19 that its security incident, initially described as affecting a "limited subset of customers," has expanded to a much broader developer community, particularly those building AI agent workflows. The attack may affect hundreds of users
GateNews4h ago
OpenAI launches GPT-5.5: 12M context, AA index tops the chart, and Terminal-Bench rewrites the agent benchmark with 82.7%
OpenAI releases GPT-5.5, focused on agent-style work and enterprise knowledge processing, and also rolls it out in ChatGPT and Codex. Key points include a 12 million token context window and an AA Intelligence Index of 60, leading Claude Opus by 4.7 and Gemini 3.1 Pro; pricing is $5 per million tokens for input and $30 per million tokens for output. Output tokens are reduced by about 40%, while the actual cost increases by about 20%.
ChainNewsAbmedia5h ago
Cluster Protocol Raises $5M to Accelerate CodeXero, Browser-Native AI IDE for EVM
Gate News message, April 23 — Cluster Protocol, an AI deeptech and Web3 infrastructure company, announced it has raised $5 million in a new funding round led by DAO5, with participation from Paper Ventures, JPEG Trading, and Mapleblock Capital, bringing total funding to $7.75 million. The capital wi
GateNews5h ago
Nvidia Expands AI Partnerships in UK, China, and Automotive Sector Amid Supply Chain Challenges
Gate News message, April 23 — Despite competition from Google and supply chain disruptions, Nvidia remains the dominant player in AI hardware. TD Cowen reaffirmed its buy rating on Nvidia on Thursday, citing the company's leadership in performance and software ecosystem breadth. The endorsement
GateNews5h ago
