Grok just broke the trend

2026-07-09 · Source: Matthew Berman · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Software Development & Engineering · Depth: Advanced, quick

Summary

Grok 4.5 has demonstrated exceptional performance on a critical agentic coding benchmark, achieving a score of 83.3. This places it five points ahead of Opus 4.8, indicating a significant advancement in AI-powered coding capabilities. Notably, this impressive result stems from a previous Grok training run, with the next-generation model already in development. This future iteration will leverage extensive cursor data obtained as part of a \$60 billion acquisition. Concurrently, Cursor is advancing its own coding model series, Composer, with Composer 2.5 recognized as a robust workhorse-class model, and Composer 3 anticipated for release soon, further intensifying innovation in this specialized AI domain.

Key takeaway

For AI Engineers evaluating coding assistant models, Grok 4.5's benchmark score of 83.3, surpassing Opus 4.8 by five points, signals a strong contender. You should monitor the upcoming next-generation Grok model, which will integrate acquired cursor data, as it could redefine coding AI capabilities. Also, consider Composer 2.5 for current workhorse coding tasks and anticipate Composer 3's release for enhanced options.

Key insights

Grok 4.5 significantly outperforms competitors on agentic coding benchmarks, with further advancements expected.

In practice

• Composer 2.5 is a robust workhorse-class coding model.

Topics

• Agentic Coding
• Grok 4.5
• Opus 4.8
• Composer Models
• AI Coding Assistants
• LLM Benchmarks

Best for: Research Scientist, AI Product Manager, Entrepreneur, AI Scientist, Machine Learning Engineer, AI Engineer

Related on AIssential

• Grok 4.5 IS REALLY GOOD! Opus & GPT Level BUT Faster, Cheaper, & Smarter! (Fully Tested) — Grok 4.5 offers a fast, cost-effective, and token-efficient AI model optimized for coding and agent workflows.
• Cursor Composer 2.5: The AI coding agent that punches way above it’s price tag — High-performance AI coding agents can be achieved at significantly lower costs by leveraging existing powerful open-source models.
• [AINews] SpaceXAI launches Grok 4.5, first Opus-class model post Cursor acquisition — Grok 4.5 delivers near-frontier AI capability with significantly improved cost-efficiency for coding and agent tasks.
• Opus 4.6, Codex 5.3, and the post-benchmark era — Real-world agentic performance and usability now outweigh traditional benchmarks for assessing new coding AI models.
• I Built an AI App in 10 Minutes with Grok 4.5 — And It Changed My Expectations — Grok 4.5 significantly reduces the time from idea to working application by generating complete architectures and handling boilerplate.
• Grok 4.2: Powerful and Fast Coding Model! Most Powerful & Intelligent Model We've Seen? (Early Test) — Grok 4.2 is an upcoming model with enhanced multimodal capabilities and efficiency, despite some early coding limitations.

See Counsel's argued verdicts on the open AI decisions leaders are weighing →

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Matthew Berman.