Grok 4.7 is xAI's cheapest model at $2/1M input tokens — 33% less than the flagship Grok 4. It is also the best capability-per-dollar pick in the lineup.
The safest xai model worth using default, the cheaper option worth trying first, and the specialist pick — before you read the detail below.
Grok 4.7 is the strongest answer here for xai model worth using — pick it when quality of output matters more than the $2.00/1M/1M input you pay for it.
Claude Sonnet 5.5 is the cheaper way in for xai model worth using, at $2.00/1M/1M input against Grok 4.7's $2.00/1M/1M.
Grok 4.5 is the fastest of these for xai model worth using — worth it when latency is what the reader notices, not the last few points of reasoning depth.
Grok 4.7 is the lowest-cost xAI model: $2/1M input, $6/1M output.
Grok 4.7 is the best capability-per-dollar pick (budget score 64/100).
Grok 4 costs 2x more on input — reserve it for work where quality is the bottleneck.
Choose Grok 4.7 for high-volume, low-stakes tasks like classification, extraction, and drafts.
Choose Grok 4.7 as the everyday default if you want one budget model.
Route only the hardest tasks to Grok 4 — a two-tier setup usually cuts spend 60–80%.
Switch the scoring lens to see whether the xai model worth using answer changes when cost, speed, or long-document depth leads the decision.
xAI / Balanced / Oct 10, 2026
Grok 4.6's price with modest coding gains and heavier token use.
Ranks models by the broadest mix of coding, writing, research, and long-context usefulness.
You need a published SWE-bench score, or predictable cost per task — independent testing found its token use roughly doubled.
Where the xai model worth using recommendation shifts once you weigh price or latency differently.
Grok 4.6's price with modest coding gains and heavier token use.
Cost-efficient coding agent — replaced by Grok 4.6 and Grok 4.7.
Finishes agent tasks in half the turns — cheap where it counts.
Retired May 15, 2026 — replaced by Grok 4.6 and Grok 4.7.
List prices and published scores — the numbers this page's pick is built from.
| Model | Input | Output | Est. month | Context | Speed | Coding | Writing | Research |
|---|---|---|---|---|---|---|---|---|
| Grok 4.7xAI | $2.00/1M | $6.00/1M | $32 | 500k tokens | Fast | 87 | 85 | 88 |
| Grok 4.5xAI | $2.00/1M | $6.00/1M | $32 | 500k tokens | Fast | 94 | 72 | 84 |
| Grok 4.6xAI | $2.00/1M | $6.00/1M | $32 | 500k tokens | Fast | 86 | 84 | 88 |
| Grok 4xAI | $3.00/1M | $15.00/1M | $60 | 256k tokens | Fast | 80 | 70 | 78 |
Scores out of 100 — how we evaluate models. “Est. month” is 10M in / 2M out at list price: a ceiling, no discounts.
Why each one is on the shortlist for xai model worth using, what it is genuinely good at, and where we would steer you away from it.
Our pick for xai model worth using. It scores 87/100 on the coding axis we weight this page by, and nothing else in this shortlist matches it on output quality.
Grok 4.7, released September 21, 2026, targets coding and professional knowledge work over multi-hour tasks, with better self-verification and long-context handling than Grok 4.6 at the same $2/$6 price.
You need a published SWE-bench score, or predictable cost per task — independent testing found its token use roughly doubled.
An incremental update. Grok 4.7 keeps Grok 4.6's price and gains a little on coding-quality tests, but independent runs found it spending far more tokens per task. Worth testing if you are already on Grok; not a reason to switch.
Full pricing, benchmark table and release notes on the Grok 4.7 page.
Here for latency: it answers fastest of anything listed for xai model worth using, at 94/100 on coding.
xAI's first coding- and agent-focused model — the first full-scale deployment of the 1.5T-parameter V9 MoE base, trained with real developer-session data from Cursor.
You need the highest solve rate per attempt or verified low hallucination — Claude Opus 5 is the safer premium pick.
The efficiency play among coding agents. Grok 4.5 wins on cost-per-solved-task and marathon sessions, not raw capability. If your agent bill is the problem, it's the answer; if quality ceiling is the problem, it isn't.
Full pricing, benchmark table and release notes on the Grok 4.5 page.
Rounds out the shortlist for xai model worth using at 86/100 on coding.
Finishes agent tasks in half the turns — cheap where it counts. Full Grok 4.6 review →
Rounds out the shortlist for xai model worth using at 80/100 on coding.
Retired May 15, 2026 — replaced by Grok 4.6 and Grok 4.7. Full Grok 4 review →
UseRightAI recommendations are based on practical decision factors people actually feel in day-to-day use.
Newsletter
We email when the xai model worth using pick changes, when one of these models moves on price, or when something new displaces the current leader.
No spam. Useful updates only. Affiliate disclosures always clearly labeled.
Grok 4.7 at $2/1M input and $6/1M output tokens. Grok 4.6's price with modest coding gains and heavier token use.
Grok 4.7 is the best capability-per-dollar pick in xAI's lineup (budget score 64/100). It handles multi-hour coding and document tasks at mid-tier pricing well — step up to Grok 4 only where quality visibly falls short.
Grok 4.7 costs $2/1M input vs $3/1M for Grok 4 — a 33% saving on input tokens.
Grok 4.7 — 500K tokens at $2/1M input. Context is where budget models are least compromised: you usually lose reasoning depth before you lose window size, so a cheap model is often a perfectly good choice for summarising or extracting from long documents.
No published SWE-bench Verified or Pro figure, so it is absent from our SWE-bench leaderboard. Trails Claude Fable 5.1 on the vendor's own coding and terminal rows (Terminal-Bench 4.0 37.6%). Avoid it if you need a published SWE-bench score, or predictable cost per task — independent testing found its token use roughly doubled.
On a moderate workload of 10M input and 2M output tokens, Grok 4.7 runs about $32.00 against $60.00 for Grok 4 — a difference of $28.00 a month at the same volume. Output tokens dominate the bill on both, so the length of the responses you generate matters far more than the length of your prompts.
Mixing is almost always cheaper for the same quality. Route high-volume, low-stakes work — classification, extraction, first drafts, routine agent steps — to Grok 4.7, and reserve Grok 4 for the calls where a wrong answer costs real time. Teams that split this way typically cut spend substantially without a quality drop anyone notices, because most tokens in a real workload are not hard problems.