slow, overengineered, bad quality code
Anthropic · Claude
Claude is mid today.
Live Claude performance tracker — real-time votes from people shipping with it every day. No synthetic benchmarks.
Your take
How is Claude for you right now?
Right-now mood
Mid
24h avg score
3.0
0 votes
Top verdict
Genius
0% of voters
Total votes
0
24h window
Claude performance window
How the last window has felt to real users.
Claude · 7d
scale: 1 (terrible) → 5 (genius)
Verdict breakdown
24h · n = 0
Claude performance: quick answers
Live answers to what people ask about Claude today.
- Is Claude dumb today?
- Claude is currently rated mid. We track Claude performance in real-time using votes from people who ship with it every day — check back as more votes come in.
- How is Claude performing today?
- Claude has a live performance score of 3.0/5 today. The top verdict is genius (0% of voters). This score updates every time someone votes, so it reflects what Claude actually feels like for real users right now — not a synthetic benchmark.
- Is Claude getting worse?
- To see whether Claude is degrading, check the 7-day trend chart above. Users often report Claude feels different after model updates — we published a deeper explainer on why AI models silently change and what to watch for.
- Where does this Claude data come from?
- Every rating is a vote from a real user who clicked Terrible, Dumb, Mid, Smart or Genius after using Claude. We capture real-time votes from people shipping with it every day, plus sentiment from Claude Code extension, upcoming Chrome/CLI extensions, and public discussions. No synthetic benchmarks.
Latest Claude reports
What people said about Claude recently.
I've found Claude Code's performance to be a bit of a mixed bag. While it's quite decent at reading files, it's surprisingly bad at knowing which ones to target. It's incredibly frustrating on larger repos because it just wanders around grepping, which burns through my context and turns. I feel like I'm just wasting time on navigation instead of the actual work.
I've finally built an MCP server for my notification infra, and it's a total game changer! I don't even have to leave my terminal to check a dashboard anymore. Now, Claude is actually writing my workflows, sending out campaigns, and reading back delivery stats. It handles personalized emails and SMS so smoothly—it feels incredibly empowering to have this much automation at my fingertips.
I'm desperately trying to figure out how to make this AI a powerhouse for building Roblox games, but I'm hitting a wall. It keeps spitting out the same repetitive slop over and over, even when I describe exactly what I want changed! It's so frustrating to burn through my daily credits only to have it get the same thing wrong four times in a row. I just need it to be smarter.
I've been trying to digitize my academic papers in Latin and Ancient Greek, but Claude keeps slamming me with these ridiculous safety guardrail blocks every few minutes. It's so frustrating because my work is purely historical and totally harmless! I feel like I’m being blocked from my own research, and the lack of response from Anthropic is honestly making me feel quite suspicious.
Missing details it normally doesn't
I am honestly so frustrated with the way Anthropic's outputs are evolving. I just ran into this response, and it's an absolute nightmare of dense, incomprehensible jargon that makes zero sense. It feels like the product is actually regressing rather than improving. Dealing with such unintelligible nonsense is incredibly demoralizing and makes the tool feel much less useful than before.
I’m trying to transition to Claude Code, but I’m struggling to find my rhythm. While my coworkers swear it’s nearly autonomous, my experience hasn't matched that hype at all. I feel stuck in this tedious cycle of generating and iterating because I can't find a proper diff view like Cursor offered. It’s honestly a bit jarring to realize I can't just trust it blindly like they seem to.