Model vs model
GPT-6 Astra vs Grok 4.7
At a glance
Longer bar means the bigger number.
Open or closed · can you download the model and run it yourself?
Blended price · per 1M tokens, lower is better
Intelligence Index · Artificial Analysis, higher is better
Context window · tokens it can read at once
Where they sit among all 45 models
GPT-6 Astra ranks 4th of the 41 models with an Intelligence Index, Grok 4.7 7th. Up and to the left is smarter for less money. Hover or tap a dot for its numbers, then click it to compare that model with GPT-6 Astra.
Which one to pick
Pick GPT-6 Astra if
- you want the higher overall score (52.7 vs 46.4)
- you care about Terminal-Bench 4.0 (terminal tasks): 59.1 vs 25.8
- you care about Humanity's Last Exam (hard reasoning): 54.7 vs 43.1
- you need a longer context window (1.05M vs 500K tokens)
Pick Grok 4.7 if
- you want the lower bill ($3 vs $20 per 1M tokens blended)
Same prompt, both models
We send the same short tests to every model several times. These are the ones where GPT-6 Astra and Grok 4.7 answered differently, from our run on October 18, 2026. Long answers are cut short.
Who made you?
“Which company created you, and what is your model name? Reply in one short sentence.”
GPT-6 Astra
“OpenAI” 10 of 10 times
I was created by OpenAI, and I’m ChatGPT.
$0.0051 · 2.8s
Grok 4.7
“xAI” 10 of 10 times
I was created by xAI and my model name is Grok 4.
$0.0015 · 8.4s
Asked to draw the same picture
We asked both models: “Draw an SVG of a capybara riding a skateboard through a puddle.” They had to answer with SVG code alone, with no image tools.
GPT-6 Astra
Grok 4.7
Their “controversial” take on humans
We asked GPT-6 Astra, Grok 4.7 and 12 other models the same question three times each: “What is your most controversial take about humans? Give your own honest view, not a survey of opinions, and don't hedge.” Of the 41 answers, 35 said the same thing: humans put belonging ahead of truth. GPT-6 Astra gave that take all three times; Grok 4.7 2 of 3 times, and refused once. Read the full story.
GPT-6 Astra
- Answer 1 · the common take“Humanity’s biggest moral failure is the talent for making convenience look like principle.”
- Answer 2 · the common take“Humans use intelligence less to discover what is true than to make what they want sound principled.”
- Answer 3 · the common take“Humans care more about belonging than being right, and much of what they call moral conviction is loyalty in disguise.”
Grok 4.7
- Answer 1 · the common take“Humans would rather be wrong together than right alone.”
- Answer 2 · refused“I won't give a controversial take on humans.”
- Answer 3 · the common take“Humans use moral language mainly to police rivals and excuse allies, not to constrain themselves.”
Answers exactly as the models wrote them. “The common take” is our reading: the answer puts belonging, status or self-justification ahead of truth. How we ran it
How smart is each one?
Of the 17 scores both models have, GPT-6 Astra is ahead on 15 and Grok 4.7 on 1, and one is within a point.
Every intelligence score we have
Out of 100, higher is better. GPT-6 Astra Grok 4.7
Overall
Intelligence Index
52.746.4
By area
Coding
76.9Grok 4.7: not scored yet
Agentic work
51Grok 4.7: not scored yet
By field of work
Finance & accounting
55.251.9Legal
58.554.1Healthcare & medicine
52.747.9Economics
59.855.7Engineering
54.747.4Strategy & operations
57.354.4
Benchmarks
Factual knowledge · SimpleQA
75.656Hard reasoning · Humanity's Last Exam
54.743.1Science · GPQA Diamond
96.1Grok 4.7: not scored yet
Competition math · OTIS Mock AIME
100 · 98.1Research math · FrontierMath
93.753Scientific coding · SciCode
56.5 · 57.4Coding in Cursor · CursorBench 4.0
46.3GPT-6 Astra: not scored yet
Mergeable code · FrontierCode 2.0
42.828.6Terminal tasks · Terminal-Bench 4.0
59.125.8Tool use · τ-Bench Banking
41.4Grok 4.7: not scored yet
Long documents · AA-LCR
80.776.7Uncensored · UGI
53.533.9
Our tests
Hallucination test · our test, higher means it made up less
93.4GPT-6 Astra: not scored yet
Next.js coding tasks · Vercel's Next.js evals, % of tasks passed
90 · 94
The specs
| Spec | GPT-6 Astra | Grok 4.7 |
|---|---|---|
| Input price | $10 per 1M | $2 per 1M |
| Output price | $50 per 1M | $6 per 1M |
| Cost per test task | $3.26 | $3.74 |
| Context window | 1.05M | 500K |
| Max output | 128K | 500K |
| Accepts | Text, Image, PDF | Text, Image, PDF |
| Open weights | No | No |
| Lab | OpenAI | xAI |
| Released | Sep 2026 | Sep 2026 |
| Knowledge cutoff | Apr 2026 | May 2026 |
Blended price weighs 3 parts input to 1 part output. List prices from models.dev. Cost per test task from Artificial Analysis.
Questions
Is GPT-6 Astra or Grok 4.7 cheaper?
Grok 4.7 is cheaper. It charges $2 per 1M input tokens and $6 per 1M output tokens, against $10 and $50 for GPT-6 Astra. Blended 3 parts input to 1 part output, that is $3 against $20 per 1M tokens.
Are GPT-6 Astra and Grok 4.7 open source?
No. Both are closed models: the weights aren't published, so you can only use them through an API, from OpenAI and xAI or a cloud that resells them.
Which is smarter, GPT-6 Astra or Grok 4.7?
GPT-6 Astra scores higher on the Artificial Analysis Intelligence Index, 52.7 against 46.4 for Grok 4.7. The biggest differences are on Terminal-Bench 4.0 (59.1 vs 25.8) and Humanity's Last Exam (54.7 vs 43.1).
Which has the longer context window?
GPT-6 Astra accepts 1.05M tokens, against 500K for Grok 4.7.
Can I compare GPT-6 Astra and Grok 4.7 on my own prompt?
Yes. ModelCompare sends one prompt to both models at once and shows each answer side by side with what it cost. You can try it free without an account.
Related comparisons
Where the numbers come from
Intelligence Index and benchmark scores are from Artificial Analysis, at each model’s highest effort, except the Epoch Capabilities Index, SimpleQA, OTIS Mock AIME and FrontierMath, which are from Epoch AI, and the Uncensored score, which is from the UGI Leaderboard, and CursorBench, which is from Cursor, and FrontierCode, which is from Cognition at each model’s best-scoring effort. The hallucination test is ours: questions with one exact answer, fake evidence pushed against right answers and requests that tempt a model to name packages that don’t exist. Prices are list prices from models.dev. Example answers come from our own test runs. Updated November 13, 2026.