AI Briefing
KO

Qwen vs Gemma Pac-Man Comparison

·2026.05.01 10:03

Key point

In a MacBook Pro M5 Max test, Gemma 4 31B outperformed in Pac-Man generation quality.

Details

Environment

On a MacBook Pro M5 Max with 64GB RAM, Pac-Man style game generation was compared using the same single prompt.

Results

  • Qwen 3.6 27B: 32 tok/s, 18 minutes 04 seconds, 33,946 tokens
  • Gemma 4 31B: 27 tok/s, 3 minutes 51 seconds, 6,209 tokens

Qwen produced a longer, more creative, and visual response, while Gemma delivered shorter, clearer code faster.

In this one-shot Pac-Man test, Gemma 4 31B came out ahead with more stable game logic, click responsiveness, and wall/ghost/particle interactions.

benchmarkqwengemmalocal-llmgamedev
Read originalShare on Threads

This summary was generated automatically by AI. Check the original for the author's claims and context. Copyright belongs to the original author.

Our guide explains how the AI works. Report summary errors, attribution issues, or removal requests via Contact.

Related

RedditMica v0.1 4B Released: Open-Source Decision Model for Agent Loops Trained for Under $30TLDR AIQwen Intelligence Launches Three Mobile AI Agents and Open Benchmark SuiteRedditUkisAI Releases Swift Model Boosting Qwen 3.8 27B Inference Speed by 1.95xHacker NewsApple Releases LensVLM-9B for Processing Compressed Text ImagesAnthropic NewsAnthropic Releases Claude Opus 5.5 with 40% Cost Reduction and Record Alignment ScoresTLDR AIQwen Releases 'RecreationWorld' for Training AI Agents to Rebuild Apps