GPT-5.6 Luna vs. GPT-6 Astra: Is a $1.20 Model Good Enough for Code Review?

SteadyVane has taken 34 readings of GPT-5.6 Luna vs. GPT-6 Astra: Is a $1.20 Model Good Enough for Code… since Sep 14, 2026: 71 points at the first reading, 165 at the latest (+94). Its biggest one-day gain was +84 on Sep 15, 2026.

open the story · Hacker News thread

Day by day

Each day's last reading. Change: against the day before (after days with no reading: since the last day that has one); the first day shows the change within that day.

On the tracker; not yet featured in a brief.

More rising on Hacker News

• AI models ran real businesses: They sent $12,431 in fake invoices, lost $3,200+992 → 101 points
• Apple's Siri AI Can Be Swapped Out for Claude, ChatGPT, Code Shows+97129 → 226 points
• GPT 5.6 Cyber+9634 → 130 points
• Autolith: A programming agent with a live runtime+9629 → 125 points
• Show HN: Ante, a coding agent in a single binary that runs offline+9572 → 167 points
• Launch HN: Bullet (YC S26) – A Faster Coding Agent+9522 → 117 points
• Show HN: Engrim – A universal, local-first SQLite memory engine for AI CLIs+856 → 91 points
• Show HN: Nari Qwen3-TTS and Qwen3-ASR – High accuracy, low latency and cost+7911 → 90 points

Everything rising on Hacker News →