xAI and Cursor launched Grok 4.5, a new 1.5T-parameter model designed for coding and AI agents, priced at $2 per 1M input tokens and $6 per 1M output tokens.
It ranks #4 on independent benchmarks and uses 60% fewer output tokens than Opus 4.8 per task, positioning it as a cost-efficient alternative to GPT-5.6 and Opus 4.8 for developers and engineering workflows.
What happened
xAI and Cursor jointly launched Grok 4.5, positioned as the first model trained specifically for coding and agents. Grok 4.5 is 3x larger than Grok 4.3 at 1.5T parameters, priced at $2 per 1M input tokens and $6 per 1M output tokens, with a 500k context window (expandable to 1M soon). It ranked #4 on Artificial Analysis Intelligence Index with a score of 54 and #4 on GDPval-AA v2 Elo at 1543.
Why it matters
Grok 4.5 targets the coding-agent workflow market dominated by Anthropic and OpenAI, offering near-Opus quality with materially better economics and speed. Its pricing undercuts GPT-5.6 ($5/$30) and Opus 4.8 ($5/$25) significantly, and efficiency metrics show average output tokens per task are over 60% lower than Opus 4.8, making it potentially attractive to developers and engineering teams seeking cost-effective frontier capabilities.
What to watch
Grok 4.5 is available immediately in Grok Build, the API, Cursor, Hermes Agent, OpenRouter, and Grok subscriptions. Cursor is offering double usage for the first week. Cache hits are discounted by 75% to $0.5 per 1M tokens, and long inputs over 200k tokens cost double.
Ask the AI about this article →
xAI's Grok 4.5 launch reflects a strategic shift toward the coding-and-agents market segment that has become central to LLM (AI that understands and generates text) competition. The body frames Grok 4.5 not as a general-purpose flagship chasing absolute benchmark supremacy, but as a specialist model optimized for efficiency and speed in engineering workflows. This positioning aligns with Elon Musk's framing of the model as "roughly comparable to Opus 4.7, but much faster," emphasizing utility to Tesla and SpaceX engineers over benchmark chasing.
The efficiency gains are substantive: independent evaluation by Artificial Analysis found that Grok 4.5 achieves near-frontier capability while using 60% fewer output tokens per task than Opus 4.8, translating to lower cost per task ($0.31–$2.59 depending on task type). The pricing undercuts both OpenAI's GPT-5.6 and Anthropic's Opus 4.8, which positions Grok 4.5 as a value proposition for cost-sensitive developer and enterprise use cases, particularly those running agents (autonomous AI systems) that incur high token consumption. The body notes that Grok 4.5 is 3x larger than Grok 4.3, a substantial architectural jump that supports its capability gains.
For example, today's edition would include:
AI-summarized, only the topics you pick — one digest a day via Email, Slack, or Discord.
Free · takes 30 seconds · unsubscribe anytimeWhat is AIToday? →
Ask AI anything about this article. Q&As are published on this page for other readers too.
Visko raised $10 million in pre-seed funding from Llama Ventures and opened public access to its first foundat…
U.S. markets ended August higher, with the S&P 500 up 2.6% and the Nasdaq up 3.9%
Neurovia AI, an Abu Dhabi-based company, is pitching Saudi security agencies software that it says can compres…
AI company Runway has unveiled Solaris, the first model in a new category it calls "Interface World Models." I…
Google's AI search gave advice to call emergency services for users alone with an African, Indian, or Pakistan…
John Deere introduced JD, a conversational AI tool that lets farmers ask open-ended questions about their hist…
