SpaceXAI Released Grok 4.5 — Opus-Class Code Model
SpaceXAI released Grok 4.5 model developed in collaboration with Cursor. The model is positioned as competitor to OpenAI and Anthropic: operates at 80 TPS, costs $2 per million input tokens and $6 per million output tokens. Ranks first on Harvey Legal Agent Benchmark.
AI-processed from MarkTechPost; edited by Hamidun News
# SpaceXAI Releases Grok 4.5: Opus-Class Coding Model
On July 14, 2026, SpaceXAI released Grok 4.5 — its first release following the company's IPO and acquisition of Cursor, a startup specializing in AI-powered coding.
Grok 4.5 Characteristics
- Opus class — comparable to the best competitor models
- Trained jointly with Cursor
- Performance: 80 TPS (transactions per second)
- First release of the unified Cursor + SpaceXAI product
Pricing
- $2 per million input tokens
- $6 per million output tokens
- Competitive pricing compared to OpenAI and Anthropic
Specialization
Grok 4.5 is positioned specifically for:
- Coding and code analysis
- Agentic tasks (autonomous systems)
- Knowledge work (working with knowledge)
- NOT for casual chat
Achievements
The model ranked first in the Harvey Legal Agent Benchmark, demonstrating its ability to work with agent systems on complex legal tasks.
What This Means
This demonstrates that SpaceXAI, with support from Elon Musk, is seriously competing in the market for advanced AI models. The focus on coding and agentic systems points to the target audience — developers and companies integrating AI into their processes.
Need AI working inside your business — not just in your newsfeed?
I build production AI for companies — custom CRM, internal tools, autonomous agents, workflow automation. Owned by you, shaped to your process, no per-seat tax. Built by Zhemal Khamidun, CPO of AlpinaGPT (AI platform, 6,000+ users).
The AI world, distilled — once a week
Seven stories that actually mattered, hand-picked. No noise, no reposts, no press releases.
Done! Check your inbox for a confirmation.