SpaceXAI launches new model, Grok 4.5

SpaceXAI announced Grok 4.5, its newest agentic‑coding model that outperforms competitors on engineering benchmarks and is priced aggressively for enterprise use.

SpaceXAI unveiled Grok 4.5, its latest agentic‑coding model, promising superior performance on engineering benchmarks while offering aggressive pricing for enterprise customers.

Key Features of Grok 4.5

Grok 4.5 builds on the architecture of its predecessor, adding a larger context window and refined code‑generation heuristics that enable more accurate completion of complex programming tasks.

The model is designed to handle multiple programming languages, with particular strength in Python, JavaScript, and Rust, according to internal testing conducted by SpaceXAI’s research team.

Performance Benchmarks

In head‑to‑head comparisons on standard engineering suites such as HumanEval and MBPP, Grok 4.5 achieved higher pass rates than rival models from major AI vendors, though exact percentages were not disclosed.

SpaceXAI highlighted faster inference times, noting that the model can generate code snippets in under a second for typical queries, a speed advantage that could benefit large‑scale development pipelines.

Pricing and Enterprise Availability

The company announced a tiered pricing structure aimed at enterprise users, with a base rate that undercuts comparable offerings from competitors. Volume discounts and custom licensing options are also available for organizations with high usage demands.

Pay‑as‑you‑go tier for small teams
Committed‑use contracts with discounted rates
On‑premises deployment for regulated industries

Early adopters are expected to integrate Grok 4.5 into internal code review tools, automated testing frameworks, and developer assistants to accelerate software delivery cycles.