Grok 4.5 undercuts the frontier on cost

XAI released Grok 4.5 on 8 July 2026, priced at $2 per million input tokens and $6 per million output tokens, making it roughly five times cheaper than competitors like Claude Fable 5 and GPT-5.5. The model ranks fourth on the Artificial Analysis Intelligence Index but offers up to 15 times lower cost per finished coding task, positioning it as a cost-effective option for coding agents and enterprise workflows.

What xAI shipped xAI released Grok 4.5 on 8 July 2026, trained on a cluster of Nvidia’s latest Blackwell-generation accelerators with a particular lean toward coding and agentic work. It is available today in three places: Grok Build xAI’s coding agent , Cursor the AI-first code editor , and the xAI API directly. Office plugins for Word, PowerPoint and Excel are also live, per the Decoder’s launch write-up https://the-decoder.com/grok-4-5-is-so-cheap-compared-to-fable-5-and-gpt-5-5-that-benchmark-gaps-may-not-matter-much/ . The story is not the benchmark ranking. It is the price. Grok 4.5 costs $2 per million input tokens — a token being the small chunk of text these models charge by — and $6 per million output tokens. Claude Fable 5, the model that actually tops the shared coding evaluations, charges $10 and $50. OpenAI’s GPT-5.5 charges $5 and $30. Grok is roughly a fifth of both on output, before tokens-per-task even comes into it. The cost compounds Grok 4.5 also uses markedly fewer output tokens per resolved coding task than its peers — the exact figures are in the box. Compound the cheap per-token price with the lean output and you land somewhere between ten and fifteen times cheaper per finished coding task against the established frontier, per CodingFleet https://codingfleet.com/blog/grok-4-5-vs-gpt-5-5/ . ≈15×cheaper per finished coding task than GPT-5.5 or Opus 4.8 — once cheap output price and lean tokens-per-task are combined. Artificial Analysis https://the-decoder.com/grok-4-5-is-so-cheap-compared-to-fable-5-and-gpt-5-5-that-benchmark-gaps-may-not-matter-much/ , the independent benchmark service that aggregates model performance across many tasks into a single score, ranks Grok 4.5 fourth on its Intelligence Index — behind Fable 5, GPT-5.5, and Opus 4.8. It gained 16 points over Grok 4.3, putting xAI close to frontier performance. It trails OpenAI and Anthropic but beats every open-weights model and Google’s Gemini. That ranking matters, but not as much as the price suggests. Coding agents have become a budget line item, not a research decision. A team running 1,000 coding tasks a day on GPT-5.5 spends around £37. The same volume on Grok 4.5 lands at about £2.50. Over a year, that gap is the difference between a junior contractor and a line on the spreadsheet you stop reading. What you can use it for today Three routes for a UK team right now: Grok Build — xAI’s coding agent think Claude Code, but with a Grok brain , free for a limited time. The lowest-friction way to try Grok 4.5 on real work today. Cursor — the AI-first code editor, with native Grok 4.5 integration on every paid plan. xAI co-trained the model with Cursor on developer workflow data, so the integration is tighter than a generic API add-on, per Digital Applied https://www.digitalapplied.com/blog/grok-4-5-vs-opus-4-8-vs-gpt-5-5-best-frontier-model-2026 . xAI API — OpenAI-compatible at the same $2/$6 pricing. The route for anything you want to script or wire into an existing toolchain. Office plugins widen the use case beyond pure coding: draft a tender response, summarise a long spreadsheet, rewrite a partner’s brief. All three of those are token-heavy on input, cheap on output — exactly Grok’s sweet spot. One UK-shaped caveat: Grok 4.5 is not yet available in the EU. xAI is targeting a mid-July launch. If your team is on UK accounts, confirm access before wiring it into a production workflow. Sources & quotes Every quotation in this article is verbatim from a named source — click any 1 to see where it came from. It's part of how we keep an AI-run newsroom honest. How we verify → /blog/how-we-keep-an-ai-newsroom-honest/