Grok 4.5 Pricing: xAI's New Coding Model Undercuts Claude Opus and GPT-5.5

xAI's Grok 4.5 pricing is the story that stands out from a crowded month of model launches: $2 per million input tokens and $6 per million output tokens, undercutting flagship rivals by well over half while landing fourth on a leading independent benchmark. Released July 8, 2026, it's the company's first model built specifically for coding and agentic work, and it arrives at a moment when nearly every AI lab is racing to cut prices rather than raise them.

What Grok 4.5 Actually Costs

According to xAI's published API pricing, Grok 4.5 runs $2 per million input tokens, $0.50 per million cached input tokens, and $6 per million output tokens, with higher effective costs for long-context or heavy tool-use requests. That's the same $2/$6 structure now common among mid-tier flagship models, but the comparison that matters is against Claude Opus 4.8 and GPT-5.5, where independent reporting puts Grok 4.5's pricing at more than 60% below both on a per-token basis. For a model xAI is positioning as Opus-class on coding benchmarks, that's an aggressive price-to-capability claim.

Built on the Cursor Acquisition

Grok 4.5 isn't a standalone release, it's the first major product to come out of SpaceX's roughly $60 billion acquisition of Anysphere, the startup behind the popular Cursor coding editor. That lineage shows in the positioning: xAI is explicitly targeting coding agents and professional knowledge work rather than general chat, and Grok 4.5 ships built into Grok Build, xAI's own coding-agent harness, alongside availability across all Cursor plans. Third-party access already spans OpenRouter, Vercel, Cloudflare, Snowflake, and Databricks Mosaic, which is a wider distribution footprint than most model launches get in their first week.

How It Stacks Up on Benchmarks

Grok 4.5 lands fourth on the Artificial Analysis Intelligence Index, ahead of every open-weight model currently ranked and notably above all Gemini models, according to benchmark aggregators tracking the release. On coding tasks specifically, independent testing has it performing roughly on par with GPT-5.5 in Codex-style evaluations at around half the per-task cost. Whether that holds up across messier, real-world codebases rather than benchmark suites is the usual caveat with any freshly launched model, benchmark parity doesn't always translate to production parity, especially on long-horizon agentic tasks where earlier Grok versions have lagged rivals.

The Catch: Not Yet in the EU

Grok 4.5 is not currently available in the EU, either through xAI's own products or the API console, with the company indicating EU availability is targeted for mid-July 2026. For a model being marketed on immediate developer adoption, that's a meaningful gap, and it puts xAI behind competitors like Anthropic and OpenAI who already serve EU customers directly. Pricing announcements rarely mention geographic restrictions up front, so this is worth flagging for anyone budgeting a coding workflow around Grok 4.5 specifically.

Why This Matters Beyond xAI

Grok 4.5's launch fits a pattern that's defined July 2026: GPT-5.6, Meta's Muse Spark 1.1, and now Grok 4.5 have all shipped with aggressive per-token pricing aimed squarely at developers, not just chat subscribers. When multiple frontier-adjacent labs are undercutting each other by 50-60% on the same axis, the practical effect is that picking one model and sticking with it gets more expensive relative to the alternative of routing work to whichever model is cheapest for a given task. That's the same logic behind bring-your-own-key tools like ByteChat, which let you hold API keys for multiple providers at once and compare them directly instead of committing to a single vendor's roadmap.

Frequently asked questions

How much does Grok 4.5 cost via the API?

Grok 4.5 is priced at $2 per million input tokens, $0.50 per million cached input tokens, and $6 per million output tokens, according to xAI's published API pricing as of July 2026.

Is Grok 4.5 cheaper than Claude Opus 4.8 or GPT-5.5?

Yes. Reporting on the July 8, 2026 launch puts Grok 4.5's pricing at over 60% below both Claude Opus 4.8 and GPT-5.5 on a per-token basis, while xAI claims comparable coding performance.

Is Grok 4.5 available everywhere?

Not yet. Grok 4.5 is not currently available in the EU through xAI's products or API console; xAI has said EU availability is targeted for mid-July 2026.

A 60% price gap against Opus-class rivals is a bold opening bid, benchmark parity under real workloads is the part still worth watching.