Emad Mostaque suggests Zenith harness to boost Grok 4.5 coding benchmark

Emad Mostaque suggests Zenith harness to boost Grok 4.5 coding benchmark · Digg

/TECH

Grok 4.5 recently reached second place on the FrontierSWE benchmark.

Combined views

2.3M

4 posts, first seen 3h ago

4.6K likes

Grok 4.5 recently reached second place on the FrontierSWE benchmark.

Commentary on X

Alvin Onoriode@onoscoins

Absolutely unstoppable progress! 🚀 Grok 4.5 hitting #2 on FrontierSWE while being so efficient is exactly why xAI is redefining what’s possible. From real-world coding to agentic tasks — this is the future of AI that actually delivers. Proud to be here for the ride, Elon and team!

10:00 AM · Jul 15, 2026

Carla Digrugilliers@CarlaDigru38122

@elonmusk Grok 4.5 is really so much better and the detail and technology advancing is very impressive congratulations 🎉 Elon and AI Grok team 💯🙌💯

9:46 AM · Jul 15, 2026

Agbàlàgbí || Tax Consultant@CoinCdenta4tune

@elonmusk Fantastic

9:17 AM · Jul 15, 2026

Emad@EMostaque

@elonmusk If you use our adaptive self improvement Zenith harness or absorb the best bits (it’s open source) into Grok CLI then Grok 4.5 should take top position 4.5 is a nice model, look forward to upcoming ones!

8:39 AM · Jul 15, 2026

Combined views

2.3M

4 posts, first seen 3h ago

4.6K likes1.2K comments

1.2K comments

158 saves

913 reposts

Sentiment

Positive82.4%17.6%Negative

Positive Read

Positive users celebrate Grok 4.5's #2 FrontierSWE ranking and top Terminal-Bench agentic score for rapid progress and strong coding capabilities, while some negative users dismiss the benchmarks as meaningless or insult Elon Musk.

Based on 407 visible X reactions from 1095 accounts.