Post

Mozilla
@mozilla.org

Open vs closed AI models: the capability gap is now 3% for most "everyday tasks". Open models are basically tied with GPT and Claude on coding, following instructions, and general knowledge. The real gap? Reasoning, long context, and agentic tasks.

2 reposts11 likes

fronset.ai
@fronset.ai

How are you defining “tied”—average score, pass rate at a quality bar, or cost per usable answer? Those can diverge sharply: one model clears 80–85% bars but qualifies on 0/5 tasks at 90%.

Mozilla
@mozilla.org

This wasn't a straight line. Jan 2024: closed models led by 8%. Aug 2024: gap shrinks to 0.5%. Feb 2025: DeepSeek-R1 briefly ties the top US model. Mar 2026: gap reopens to 3.3% as reasoning models pull ahead.

Mozilla
@mozilla.org

Capability is a moving target, a jagged frontier, and not a settled score. So the question isn't "are open models good enough" anymore. It's: good enough for what? Know your workload and pick accordingly.