Grok 4.5: The New Flagship Model

Updated November 2, 2026 — flagship model. Grok 4.5 is no longer the most capable model: Grok 4.6 (August 12, 2026) and then Grok 4.7 (grok-4.7, September 21, 2026) replaced it, with 500,000 context tokens, reasoning effort from low to xhigh, and the same tiered pricing ($2 / $6 under 200,000 prompt tokens, $4 / $12 above). grok-4.5 remains available; for a new project, start with grok-4.7 (course Grok in 2026, lesson 9).

On July 16, 2026, xAI announced Grok 4.5 and presented it as its smartest model, “for code and everything else.” Behind the tagline lie three concrete promises: agentic tool calls—the model knows how to chain actions in a workflow, not just answer—minimal hallucinations, and configurable reasoning, meaning you decide how much thinking effort the model invests before responding. This lesson details what the official documentation says about the model, and especially the pricing mechanism that should guide how you use it.

Technical Specifications

Here are the characteristics as published on docs.x.ai, captured on July 28, 2026:

Tiered Pricing: The Subtlety That Traps Budgets

Look again at the four price lines in the table, because that’s where your bill is decided. Grok 4.5 doesn’t have one price: it has two, and the second is exactly double the first. As long as the sent context stays under 200,000 tokens, you pay $2 per million input tokens and $6 for output. Once the request crosses this threshold, the whole thing switches to $4 and $12. The direct consequence: an application that systematically loads large contexts—out of convenience, because “putting everything in” is simpler than sorting—will pay the high rate permanently, without anyone consciously deciding to.

Take a typical case to make the mechanism tangible. A team connects Grok 4.5 to its document base and, for every question, injects all relevant documents “just in case.” Every request comfortably exceeds 200,000 tokens: the team pays $4 / $12 on the entire volume. The same team, filtering upstream to load only the truly useful passages, brings most of its requests under the threshold and drops back to $2 / $6—half the cost, for often better answers since the model receives less noise.

In practice, this requires two reflexes. First, measure the actual distribution of context sizes your application sends: without this measurement, you don’t know what rate you’re effectively paying. Second, load the large context only when the task truly justifies it—the 200,000-token threshold becomes a design parameter for your application, not a billing detail you discover at the end of the month.

Positioning in the Lineup

Grok 4.5 is the flagship, but it doesn’t replace the entire lineup: the documentation still lists grok-4.3, with its one-million-token context, and the grok-4.20 variants that the next lesson details. This is precisely where an apparent paradox appears that you need to keep in mind: the new flagship model has a smaller context (500k) than the previous generation models (1M). This isn’t an anomaly; it’s a trade-off. If your primary need is to ingest huge volumes—analyzing an entire corpus, loading a massive codebase in one go—the 4.x lineup remains relevant and even preferable. If your primary need is reasoning capability and code, Grok 4.5 is the choice xAI highlights. Model selection becomes a real decision again, based on what you prioritize: capability or volume.

Availability

For consumers, Grok 4.5 has been deployed on iOS, Android, web, and X since July 22. For developers, API access happens via the usual console, under the identifier grok-4.5. Note finally that it has become the default model for Grok Build, the code agent we will study in a few lessons: if you have used Grok Build since July 16, it is already Grok 4.5 working for you.

Official source: docs.x.ai/developers/models and the announcement “Introducing Grok 4.5” from July 16, 2026

Prices captured on August 5, 2026 — prices change regularly: before any budget calculation, check the current grid on the official xAI models and pricing page.

Key Takeaways

• grok-4.5: 500k context, configurable reasoning
• Tiered pricing: $2 / $6 under 200k context, $4 / $12 above
• 4.x lineup remains available, with larger context (1M)
• Choice: capability (4.5) vs very large context (4.3 / 4.20)