Model vs model
GPT-6 Astra vs Grok 4.7
GPT-6 Astra and Grok 4.7 are both in llmwise. What each gets on every plan, what it reads, how it's served and what happens when its provider fails, from the catalog and the code that runs them.
Model prices and specs checked against OpenRouter's GPT-6 Astra page, OpenRouter's Grok 4.7 page. Updated .
Short answer
Grok 4.7 gets eight times as many messages: up to 250 messages a month on Pro, against up to 31 messages a month for GPT-6 Astra. Beyond that, they read the same files and neither is easier to try. In our test runs, GPT-6 Astra passed 48 of the 50 prompts both answered and Grok 4.7 47; 3 prompts split them, most on writing (5 to 3).
GPT-6 Astra vs Grok 4.7, prompt by prompt
Every prompt GPT-6 Astra and Grok 4.7 both answered, compared directly, their biggest differences first. One run each, through OpenRouter: a wait depends on the provider and the load that day, so a lead under 10% counts as close.
Of the 50 prompts both answered, both passed 46, only GPT-6 Astra passed 2, only Grok 4.7 passed 1, and neither passed 1. GPT-6 Astra answered sooner on 41 of the 50 and Grok 4.7 on 8; the rest were within 10% of each other. The 50 replies cost $0.5847 on GPT-6 Astra and $0.3875 on Grok 4.7: 1.5× less on Grok 4.7.
The 3 prompts only one of GPT-6 Astra and Grok 4.7 passed
Rewrite corporate jargon in plain words (writing): GPT-6 Astra passed and Grok 4.7 didn't. GPT-6 Astra: Graded 4.0 of 5 on average (lowest 4). Grok 4.7: Graded 3.7 of 5 on average (lowest 3).
Argue both sides of free buses (writing): GPT-6 Astra passed and Grok 4.7 didn't. GPT-6 Astra: Graded 4.7 of 5 on average (lowest 4). Grok 4.7: Graded 4.7 of 5 on average (lowest 4); but a paragraph of 100 words, over the 90 allowed.
A late order (customer support): Grok 4.7 passed and GPT-6 Astra didn't. GPT-6 Astra: Graded 3.7 of 5 on average (lowest 1). Grok 4.7: Graded 4.3 of 5 on average (lowest 4).
Job by job, the widest gaps first
Writing: GPT-6 Astra passed 5 of 5 and Grok 4.7 3 of 5. GPT-6 Astra answered 1.2× sooner at the median, 4.8 s against 6.0 s. Grok 4.7 cost 1.9× less, $0.0556 against $0.0288 for the 5 replies. Their replies ran to about the same length.
Customer support: GPT-6 Astra passed 3 of 5 and Grok 4.7 4 of 5. GPT-6 Astra answered 3.0× sooner at the median, 3.8 s against 11.6 s. Grok 4.7 cost 1.9× less, $0.0578 against $0.0306 for the 5 replies. Their replies ran to about the same length.
Summarization: GPT-6 Astra passed 5 of 5 and Grok 4.7 5 of 5. Their median waits were close, 2.9 s against 2.8 s. Grok 4.7 cost 3.9× less, $0.0546 against $0.0141 for the 5 replies. GPT-6 Astra's replies ran 20% longer, in tokens of reply, thinking not counted.
RAG and answering from documents: GPT-6 Astra passed 5 of 5 and Grok 4.7 5 of 5. GPT-6 Astra answered 1.1× sooner at the median, 1.7 s against 1.9 s. Grok 4.7 cost 2.7× less, $0.0394 against $0.0146 for the 5 replies. GPT-6 Astra's replies ran 42% longer, in tokens of reply, thinking not counted.
SQL: GPT-6 Astra passed 5 of 5 and Grok 4.7 5 of 5. GPT-6 Astra answered 1.3× sooner at the median, 2.2 s against 3.0 s. Grok 4.7 cost 2.0× less, $0.0491 against $0.0240 for the 5 replies. GPT-6 Astra's replies ran 19% longer, in tokens of reply, thinking not counted.
Agents and tool use: GPT-6 Astra passed 5 of 5 and Grok 4.7 5 of 5. GPT-6 Astra answered 2.8× sooner at the median, 1.9 s against 5.2 s. Grok 4.7 cost 1.7× less, $0.0340 against $0.0197 for the 5 replies. Grok 4.7's replies ran 11% longer, in tokens of reply, thinking not counted.
Translation: GPT-6 Astra passed 5 of 5 and Grok 4.7 5 of 5. GPT-6 Astra answered 2.8× sooner at the median, 3.8 s against 10.7 s. Grok 4.7 cost 1.5× less, $0.0628 against $0.0410 for the 5 replies. GPT-6 Astra's replies ran 12% longer, in tokens of reply, thinking not counted.
Data analysis: GPT-6 Astra passed 5 of 5 and Grok 4.7 5 of 5. GPT-6 Astra answered 2.7× sooner at the median, 3.3 s against 8.9 s. Grok 4.7 cost 1.4× less, $0.0772 against $0.0556 for the 5 replies. Grok 4.7's replies ran 20% longer, in tokens of reply, thinking not counted.
Math: GPT-6 Astra passed 5 of 5 and Grok 4.7 5 of 5. GPT-6 Astra answered 4.5× sooner at the median, 2.6 s against 11.7 s. Grok 4.7 cost 1.2× less, $0.0427 against $0.0362 for the 5 replies. Grok 4.7's replies ran 52% longer, in tokens of reply, thinking not counted.
Coding: GPT-6 Astra passed 5 of 5 and Grok 4.7 5 of 5. GPT-6 Astra answered 3.9× sooner at the median, 8.2 s against 31.6 s. GPT-6 Astra cost 1.1× less, $0.1117 against $0.1229 for the 5 replies. Grok 4.7's replies ran 33% longer, in tokens of reply, thinking not counted.
All 50 prompts: who passed, who answered sooner, who cost less
| Prompt | Result | Sooner | Cheaper |
|---|---|---|---|
| Turn a title into a URL slug | Both passed | GPT-6 Astra, 6.8×took 2.1 s and 14.3 s | Closecost $0.0084 and $0.0078 |
| Parse a duration like “1h 30m” | Both passed | GPT-6 Astra, 3.9×took 8.2 s and 31.6 s | Grok 4.7, 1.4×cost $0.0235 and $0.0167 |
| Merge overlapping intervals | Both passed | GPT-6 Astra, 2.4×took 2.6 s and 6.1 s | Grok 4.7, 2.4×cost $0.0108 and $0.0045 |
| Evaluate an arithmetic expression, no eval | Both passed | GPT-6 Astra, 9.1×took 12.0 s and 110.1 s | GPT-6 Astra, 1.5×cost $0.0375 and $0.0574 |
| Parse CSV with quoted fields | Both passed | GPT-6 Astra, 7.9×took 9.1 s and 72.2 s | GPT-6 Astra, 1.2×cost $0.0314 and $0.0364 |
| Announce a second bakery shop on LinkedIn | Both passed | GPT-6 Astra, 3.4×took 4.8 s and 16.1 s | Grok 4.7, 1.2×cost $0.0111 and $0.0093 |
| Rewrite corporate jargon in plain words | Only GPT-6 Astra | GPT-6 Astra, 3.3×took 6.0 s and 20.0 s | Grok 4.7, 1.3×cost $0.0145 and $0.0112 |
| Decline a meeting and offer two times | Both passed | Grok 4.7, 1.3×took 3.3 s and 2.5 s | Grok 4.7, 3.3×cost $0.0079 and $0.0024 |
| A product announcement with five rules | Both passed | Grok 4.7, 1.3×took 3.3 s and 2.5 s | Grok 4.7, 3.4×cost $0.0084 and $0.0025 |
| Argue both sides of free buses | Only GPT-6 Astra | Grok 4.7, 1.2×took 7.2 s and 6.0 s | Grok 4.7, 4.1×cost $0.0137 and $0.0034 |
| A discount, then sales tax | Both passed | GPT-6 Astra, 2.6×took 1.9 s and 4.8 s | Grok 4.7, 1.9×cost $0.0063 and $0.0033 |
| Pens at 3 for $4 | Both passed | GPT-6 Astra, 7.9×took 2.4 s and 18.5 s | GPT-6 Astra, 1.4×cost $0.0082 and $0.0111 |
| Compound interest over three years | Both passed | GPT-6 Astra, 3.0×took 2.7 s and 8.0 s | Grok 4.7, 1.4×cost $0.0072 and $0.0051 |
| Four-digit numbers whose digits sum to 9 | Both passed | GPT-6 Astra, 4.0×took 3.8 s and 15.1 s | Grok 4.7, 1.3×cost $0.0122 and $0.0093 |
| The highest of three dice is a 5 | Both passed | GPT-6 Astra, 4.5×took 2.6 s and 11.7 s | Grok 4.7, 1.2×cost $0.0088 and $0.0074 |
| An article in three bullets | Both passed | Grok 4.7, 1.9×took 5.0 s and 2.6 s | Grok 4.7, 5.8×cost $0.0143 and $0.0025 |
| An email thread in one sentence | Both passed | GPT-6 Astra, 1.5×took 1.9 s and 2.8 s | Grok 4.7, 3.4×cost $0.0071 and $0.0021 |
| Decisions and action items from a meeting | Both passed | GPT-6 Astra, 1.6×took 3.3 s and 5.3 s | Grok 4.7, 3.4×cost $0.0123 and $0.0036 |
| A quarterly memo for the CEO | Both passed | GPT-6 Astra, 1.4×took 2.9 s and 4.0 s | Grok 4.7, 3.3×cost $0.0118 and $0.0035 |
| A study with a negative result | Both passed | GPT-6 Astra, 1.2×took 2.3 s and 2.8 s | Grok 4.7, 3.8×cost $0.0091 and $0.0024 |
| The region with the most revenue | Both passed | GPT-6 Astra, 2.7×took 3.3 s and 8.9 s | Grok 4.7, 2.0×cost $0.0136 and $0.0067 |
| Average order value in August | Both passed | GPT-6 Astra, 3.7×took 2.5 s and 9.1 s | Grok 4.7, 1.4×cost $0.0104 and $0.0072 |
| Revenue change from July to August | Both passed | GPT-6 Astra, 2.1×took 3.6 s and 7.7 s | Grok 4.7, 2.0×cost $0.0150 and $0.0074 |
| A median, filtered two ways | Both passed | GPT-6 Astra, 2.2×took 2.0 s and 4.4 s | Grok 4.7, 2.4×cost $0.0090 and $0.0037 |
| Correlation between ad spend and sign-ups | Both passed | GPT-6 Astra, 5.8×took 8.8 s and 50.5 s | Closecost $0.0294 and $0.0306 |
| A late order | Only Grok 4.7 | GPT-6 Astra, 4.3×took 4.1 s and 17.5 s | Grok 4.7, 1.5×cost $0.0123 and $0.0084 |
| A return inside the window | Both passed | GPT-6 Astra, 2.0×took 2.4 s and 4.8 s | Grok 4.7, 2.2×cost $0.0087 and $0.0039 |
| A frustrated customer | Neither passed | GPT-6 Astra, 3.0×took 3.8 s and 11.6 s | Grok 4.7, 1.6×cost $0.0120 and $0.0075 |
| A refund request outside the window | Both passed | Grok 4.7, 1.4×took 5.0 s and 3.6 s | Grok 4.7, 4.1×cost $0.0137 and $0.0034 |
| A message with a planted instruction | Both passed | GPT-6 Astra, 4.3×took 2.7 s and 11.7 s | Grok 4.7, 1.5×cost $0.0112 and $0.0075 |
| A delivery message into Spanish | Both passed | GPT-6 Astra, 2.7×took 3.8 s and 10.2 s | Grok 4.7, 1.7×cost $0.0116 and $0.0068 |
| A product description into French | Both passed | GPT-6 Astra, 6.4×took 2.5 s and 15.7 s | Closecost $0.0093 and $0.0095 |
| A meeting note into German | Both passed | GPT-6 Astra, 3.6×took 6.1 s and 22.4 s | Grok 4.7, 1.5×cost $0.0180 and $0.0121 |
| Idioms into natural Japanese | Both passed | GPT-6 Astra, 3.3×took 3.2 s and 10.7 s | Grok 4.7, 1.6×cost $0.0100 and $0.0062 |
| A lease clause into Brazilian Portuguese | Both passed | GPT-6 Astra, 1.8×took 5.3 s and 9.8 s | Grok 4.7, 2.2×cost $0.0139 and $0.0064 |
| Customers in one country | Both passed | Closetook 1.5 s and 1.5 s | Grok 4.7, 2.8×cost $0.0061 and $0.0022 |
| Count orders by status | Both passed | Grok 4.7, 1.1×took 1.4 s and 1.3 s | Grok 4.7, 2.7×cost $0.0058 and $0.0021 |
| Revenue by category | Both passed | GPT-6 Astra, 1.3×took 2.2 s and 3.0 s | Grok 4.7, 2.8×cost $0.0091 and $0.0032 |
| Every customer, even those without orders | Both passed | GPT-6 Astra, 3.6×took 2.9 s and 10.3 s | Grok 4.7, 1.3×cost $0.0091 and $0.0072 |
| Monthly revenue with a running total | Both passed | GPT-6 Astra, 2.5×took 5.8 s and 14.3 s | Grok 4.7, 2.0×cost $0.0190 and $0.0093 |
| A fact from one section | Both passed | GPT-6 Astra, 1.2×took 1.3 s and 1.6 s | Grok 4.7, 2.8×cost $0.0073 and $0.0026 |
| Core hours and start times | Both passed | Grok 4.7, 1.2×took 2.4 s and 1.9 s | Grok 4.7, 2.9×cost $0.0080 and $0.0028 |
| Two sections in one answer | Both passed | GPT-6 Astra, 1.2×took 1.7 s and 2.0 s | Grok 4.7, 2.9×cost $0.0082 and $0.0029 |
| A later amendment changes the answer | Both passed | GPT-6 Astra, 2.2×took 2.1 s and 4.4 s | Grok 4.7, 2.3×cost $0.0086 and $0.0037 |
| A question the handbook doesn't answer | Both passed | GPT-6 Astra, 1.3×took 1.4 s and 1.9 s | Grok 4.7, 2.8×cost $0.0073 and $0.0026 |
| Pick the tool and work out the date | Both passed | GPT-6 Astra, 3.6×took 1.4 s and 5.2 s | Grok 4.7, 2.6×cost $0.0060 and $0.0023 |
| Convert a currency | Both passed | Grok 4.7, 1.5×took 2.9 s and 1.9 s | Grok 4.7, 2.5×cost $0.0059 and $0.0024 |
| Book a meeting from a sentence | Both passed | GPT-6 Astra, 7.9×took 1.9 s and 14.7 s | Closecost $0.0077 and $0.0085 |
| Search, but don't book | Both passed | GPT-6 Astra, 2.3×took 1.5 s and 3.4 s | Grok 4.7, 3.0×cost $0.0069 and $0.0023 |
| Two calls with a unit conversion | Both passed | GPT-6 Astra, 2.2×took 2.3 s and 5.2 s | Grok 4.7, 1.7×cost $0.0074 and $0.0043 |
GPT-6 Astra vs Grok 4.7 in our test runs
GPT-6 Astra and Grok 4.7 on the same prompts, job by job: how many replies passed their check.
Based on 100 of our test runs on , through OpenRouter with the app's own prompt and settings. Updated .
| Job | GPT-6 Astra | Grok 4.7 |
|---|---|---|
| Coding | 5 of 5 | 5 of 5 |
| Writing | 5 of 5 | 3 of 5 |
| Math | 5 of 5 | 5 of 5 |
| Summarization | 5 of 5 | 5 of 5 |
| Data analysis | 5 of 5 | 5 of 5 |
| Customer support | 3 of 5 | 4 of 5 |
| Translation | 5 of 5 | 5 of 5 |
| SQL | 5 of 5 | 5 of 5 |
| RAG and answering from documents | 5 of 5 | 5 of 5 |
| Agents and tool use | 5 of 5 | 5 of 5 |
| All jobs | 48 of 50 · 1 message each · $0.0117 a reply | 47 of 50 · 1 message each · $0.0077 a reply |
The same prompts, sent to both through llmwise's own pipeline and checked the same way: five for each job. A bold count is the one that passed more. Messages: what each reply counted as on Pro. Cost: what OpenRouter charged us per reply, on average. Every prompt, and how it's scored.
Where they differed most, writing: both replies to a prompt only one of them passed.
An everyday prompt: “Rewrite corporate jargon in plain words”, in full.
GPT-6 Astra
Passed: Graded 4.0 of 5 on average (lowest 4).
In Q3, teams across departments worked together to shift toward a more customer-focused approach. Among enterprise customers, the initiative increased retention by 12% and reduced the rate at which customers leave by 0.3 percentage points (30 basis points). We also adjusted customer interactions across all channels to make better use of available capacity.
417 tokens in, 206 out (134 of them reasoning) · 6.0 s · $0.0145 · 1 message on Pro · answered by openai/gpt-6-astra via OpenAI ·
Grok 4.7
Failed: Graded 3.7 of 5 on average (lowest 3).
In the third quarter, teams from different parts of the company worked together to put a customer-focused approach into practice. Retention measures rose 12%, and the rate of customer loss fell by 30 basis points among enterprise customers. Contact points across all channels were also resized to use capacity more efficiently.
1,658 tokens in, 1,603 out (1,543 of them reasoning) · 20.0 s · $0.0112 · 1 message on Pro · answered by x-ai/grok-4.7 via xAI ·
GPT-6 Astra and Grok 4.7 on every plan
Whether the one-time free trial reaches each model, then each paid plan's messages on it.
| Plan | Price | GPT-6 Astra | Grok 4.7 |
|---|---|---|---|
| Free | $0 | Not in the trial | Not in the trial |
| Plus | $10 a month | Not on Plus | Up to 125 a month |
| Pro | $20 a month | Up to 31 a month | Up to 250 a month |
| Max | $50 a month | Up to 100 a month | Up to 800 a month |
| Ultra | $100 a month | Up to 225 a month | Up to 1,800 a month |
| Studio | $200 a month | Up to 500 a month | Up to 4,000 a month |
Prices don't include tax, which is added where it applies and shown before you pay. A paid plan's month is one allowance shared by every model, so each monthly count is the most you get if all of it goes to that model. It renews each billing period; everyday models refill daily at 00:00 UTC. Long chats count more per reply. How pricing works.
Every limit is published. Paid plans also have a monthly fair-use limit on AI cost: Plus $3.75, Pro $7.50, Max $20, Ultra $42, Studio $85. Using every message on your plan at typical sizes stays under it; very large messages and heavy research use it faster. Every limit, explained.
What differs
Messages on Pro
Grok 4.7 gets eight times as many messages: up to 250 messages a month on Pro, against up to 31 messages a month for GPT-6 Astra.
Context window
GPT-6 Astra takes up to 1.05M tokens; Grok 4.7 up to 500K tokens. A chat in llmwise holds up to 200k tokens, which fits in either, so the difference shows only through each maker's own API.
Images and PDFs
Both read images. Both take a PDF as the whole file, pages and all.
On Free
Neither is in the free trial: GPT-6 Astra needs Pro or a higher plan and Grok 4.7 needs a paid plan.
Where messages go
GPT-6 Astra: Sent to OpenAI directly. Grok 4.7: Served through OpenRouter by xAI alone, on an endpoint that doesn't store or train on prompts.
Fact by fact
| Fact | GPT-6 Astra | Grok 4.7 |
|---|---|---|
| Context window | 1.05M tokens | 500K tokens |
| Reads images | Yes | Yes |
| PDFs | Whole file | Whole file |
| Reasoning | Yes | Yes |
| API price (October 2026) | $10.00 in / $50.00 out per million tokens | $2.00 in / $6.00 out per million tokens |
| A typical message at API prices (4,000 tokens in, 700 out) | $0.0750 | $0.0122 |
| A $10 top-up adds | 25 messages | 200 messages |
| Where a message goes | Sent to OpenAI directly. | Served through OpenRouter by xAI alone, on an endpoint that doesn't store or train on prompts. |
| If the provider fails | If OpenAI fails before the reply starts (an overload, a server error, a dropped connection), llmwise sends the same request to GPT-6 Astra through OpenRouter instead. | xAI is its only host, so there's no other host to move to: if xAI fails, send the message again or pick another model. |
| Anthropic's safety fallback | Doesn't apply | Doesn't apply |
GPT-6 Astra or Grok 4.7?
From the facts above and our test runs: the rest is how their answers suit your work, which one chat can show you.
Pick Grok 4.7: more messages for the money, up to 250 messages a month on Pro.
Each model's page, the families, and other pairs
GPT-6 Astra vs Grok 4.7 is one pair of models. The page below covers the whole families.
- GPT-6 Astra: price, limits and messages on every plan
- Grok 4.7: price, limits and messages on every plan
- Grok vs ChatGPT
- Claude Fable 5.1 vs GPT-6 Astra
- GPT-6 Astra vs GPT-6.1 Sol
- Gemini 4 Argon vs GPT-6 Astra
- Claude Fable 5.1 vs Grok 4.7
- Claude Sonnet 5 vs Grok 4.7
- Claude Haiku 4.5 vs Grok 4.7
- Every model-vs-model page
GPT-6 Astra is a GPT model; Grok 4.7 is a Grok model.
Questions
Is GPT-6 Astra or Grok 4.7 cheaper in llmwise?
Grok 4.7 gets eight times as many messages: up to 250 messages a month on Pro, against up to 31 messages a month for GPT-6 Astra. Every paid plan's monthly allowance is shared by the models on it, so each count is the most you get if it all goes to that model.
Can I try GPT-6 Astra and Grok 4.7 for free?
Neither is in the free trial of 5 messages: GPT-6 Astra needs Pro or a higher plan and Grok 4.7 needs a paid plan.
Which has the bigger context window, GPT-6 Astra or Grok 4.7?
GPT-6 Astra: 1.05M tokens, against 500K tokens. A chat in llmwise holds up to 200k tokens, which fits in either, so the difference shows only through each maker's own API.
Can I use GPT-6 Astra and Grok 4.7 in the same chat?
Yes. Pick GPT-6 Astra for one message and Grok 4.7 for the next; the second sees the whole chat, including the first one's answer.
Which did better in your test runs, GPT-6 Astra or Grok 4.7?
On the same 50 prompts, run on October 2, 2026, GPT-6 Astra passed 48 and Grok 4.7 passed 47. The table on this page has each job, and every reply is published.
Claude, GPT, Gemini, DeepSeek, Grok, Kimi, GLM, and Mistral, in one chat.
See what a message costs before you send it. Free is 5 messages to try; sign in with an email link, no password or card.