Claude.ai, API, and Claude Code Experiencing Elevated Errors
Claude.ai, the Claude API (api.anthropic.com), and the Claude Code tool are currently experiencing elevated errors. The incident was identified on 2026-04-15 at 14:53 UTC.
Current Status
- Claude.ai Web Interface: Outage
- platform.claude.com Developer Console: Outage
- Claude API (api.anthropic.com): Operational (recovered as of 16:01 UTC)
- Claude Code CLI Tool: Degraded
- Claude Cowork Service: Operational
- Claude for Government: Operational
Specific Issues
Claude.ai and the Platform are down. Login for Claude Code does not work via Claude.ai. However, Claude Code users who are already logged in are still able to use it. The Claude API has fully recovered as of 8:01 PT / 16:01 UTC. Anthropic states the issue has been identified and a fix is being implemented.
Recent History
This follows several recent incidents:
- 2026-04-14: Degraded service on usage and analytics admin API endpoints, resolved.
- 2026-04-13: Claude.ai down, affecting login for Claude.ai and Claude Code between 15:31 and 16:19 UTC, resolved.
- 2026-04-11: Email login broken for approximately one hour, resolved.
- 2026-04-10: Elevated errors on requests to Claude models (non-Opus models affected), resolved.
For developers relying on Claude's services, this means the web interface and developer console are currently unavailable, API calls are functional, and Claude Code access is limited to existing sessions. Monitor the status page for updates on the fix implementation.
📖 Read the full source: HN AI Agents
👀 See Also
Open-source models match or beat Claude Opus 4.6 on benchmarks
DeepSeek V3.2, DeepSeek R1, Kimi K2.5, and MiniMax M2.5 outperform Claude Opus 4.6 on 4 out of 5 major benchmarks including MMLU-Pro, speed, tool use, and reasoning, while being significantly cheaper.
Trading Strategy Benchmark: Cheaper AI Models Outperform Claude Opus 4.6
A benchmark tested 10 LLMs on developing trading strategies, with cheaper models like Minimax 2.5 and Gemini 3.1 outperforming Claude Opus 4.6 despite its 10x higher cost. The experiment was run three times with consistent results.
Deterministic vs Probabilistic Code Generation: Why Bun's Vibe-Coded Rust Conversion Raises Red Flags
Noah Hall argues vibe-coded 1M-line repo changes (like Bun's Zig-to-Rust) are dangerous. Contrasts deterministic transpilers vs. probabilistic LLM output. Tests aren't enough.
Qwen3.5-122B on Blackwell SM120: fp8 KV Cache Corruption Issue and Performance Findings
Testing Qwen3.5-122B on 8x RTX PRO 6000 Blackwell hardware revealed that fp8_e4m3 KV cache silently produces corrupt output without errors, requiring bf16 KV cache instead. MTP optimization provided a 2.75x single-request speedup while DeltaNet constraints blocked other optimizations.