SpaceXAI has released Grok 4.5, positioning it as their most capable model for coding, agentic tasks, and knowledge work, trained in collaboration with Cursor. While the announcement highlights strong performance in engineering benchmarks like DeepSWE and Terminal Bench, independent evaluations suggest it trails behind competitors such as Fable and GPT 5.5 in several key metrics.
Background
The release occurs in a highly competitive landscape where major tech firms are rapidly iterating on AI models specifically optimized for software development and autonomous agent workflows.
- Source
- hackernews
- Published
- Jul 9, 2026 at 02:00 AM
- Score
- 7.0 / 10