Previous xAI model, the default sub-agent model in Cursor, which the creator liked for being capable and fast. It was used to regenerate SFT trajectories to train Grok 4.6.
Grok 4.5 at a glance
- What is Grok 4.5?
- Grok 4.5 is a model from xAI. xAI's Grok 4.5 language model, a cheaper, fast coding-capable model available via the xAI API and in Cursor.
- Who makes Grok 4.5?
- Grok 4.5 is made by xAI.
- Does Grok 4.5 work in the EU?
- Available in the EU. Pricing page notes prices exclude VAT, and the Enterprise tier offers data residency options and a GDPR compliance FAQ; no EU exclusion was found. Grok 4.5 itself is not explicitly named in the EU material. Last verified 14 Nov 2026.
- How much does Grok 4.5 cost?
- API pricing. Usage-based API pricing; also included in Grok plans (Free limited, Plus, Premium, Super, Ultra, Teams, Enterprise) with shared usage pools. Per-token prices not given in the provided material. Licence: Proprietary.
- Where was Grok 4.5 shown?
- In 5 creator videos by Matthew Berman, AI Search, Matt Wolfe. Most recently by Matthew Berman on 13 Aug 2026, at 1:34 in “xAI actually did it... (Grok 4.6)”.
Sources: the vendor's own pages (linked above) and creator videos at the linked timestamps. EU status is information, not legal advice; see how we check.
Grok 4.5 is a language model from xAI (the company now presenting as SpaceXSI on x.ai). Per YouTube creator commentary, it is a previous-generation Grok model that has since been succeeded by newer versions (Grok 4.6 and 4.7 are referenced). Creators describe it as capable and fast, and it was the default sub-agent model in Cursor.
It is mainly discussed as a lower-cost executor for writing code once a plan exists, with a cost of about 50 cents per task on one cost-per-task chart. One creator said it was used to regenerate SFT trajectories for training Grok 4.6. These details come from creator commentary, not vendor documentation, and the provided pages give no specs such as context window or per-token pricing.
Access is through the xAI API (usage-based pricing, Python SDK), through Grok subscription plans that include a shared AI usage pool, and through Cursor. The weights are not released. Exact release date and benchmark details are not confirmed in the provided material.
- Fast, low-cost model suited to executing coding tasks from a plan
- Default sub-agent model in Cursor
- Available through the xAI API and Grok subscription plans
- About $0.50 per task on a cited cost-per-task chart
- Succeeded by newer Grok 4.6/4.7 models
As seen on YouTube
Watch the moment: the exact second top creators show it. Videos load only when you press play, via youtube-nocookie.com.
Shown on the cost-per-task chart at about 50 cents per task with similar intelligence to Luna.
Cheaper model recommended as an executor for writing code once a plan exists.
xAI's model for agentic coding and knowledge work, trained alongside Cursor, with a 500K context window. It trails GPT 5.5 and Fable on benchmarks but is fast (about 80 tok/s), cheap and token-efficient.
xAI's new model, competitive on Deep SWE and Terminal Bench but behind GPT-5.6. Available for free with a Grok CLI, and the API costs $2 input and $6 output per million tokens. The creator tests it on Busey bench and a website prompt.
Similar AI tools
Closest neighbours by meaning, not just by category.
Grok 4.7
Model · xAI
xAI's Grok 4.7 language model for coding and knowledge work, available via API, Cursor and Grok Build.
Grok 4.6
Model · xAI
xAI's Grok 4.6 closed LLM for reasoning and coding, available via API, Cursor and Grok Build.
Grok Build
App · xAI
xAI's coding agent for Grok, bundled with Grok subscriptions and working alongside Cursor, with background memory.
GLM 5.2
Model · Zhipu AI
Zhipu AI's GLM 5.2 large language model family, with Fast and max variants, available via Z.ai chat and API.
GPT-5.5
Model · OpenAI
OpenAI language model available through ChatGPT, Codex and the API, used for coding, review and general tasks.
GPT-5.4
Model · OpenAI
OpenAI frontier language model used for coding, reasoning and agentic tasks, available via ChatGPT and the API.