Compare / Grok 4.6 vs Grok 4.7

Grok 4.6 vs Grok 4.7

Which is cheaper, which is more capable, and which does better at each kind of work, from 19 results on 3 sources that measured both models.

xAIGrok 4.6 Released 12 Aug 2026 · $2.00 in · $6.00 out
xAIGrok 4.7 Released 21 Sep 2026 · $2.00 in · $6.00 out
Grok 4.6 wins 59 too close to callGrok 4.7 wins 5

Of the 19 results both models have, the better value on each. Results whose reported ranges overlap are too close to call.

Results as of 19 November 2026 (catalogue release 2026-11-19-3608133e8331).

Grok 4.6 or Grok 4.7? The short answer

Grok 4.7 is the newer version at about the same price, and it scores higher (Intelligence Index 46.4 against 44.3): upgrade.

  • They cost about the same: $3.00 and $3.00 per million tokens (three input to one output).
  • Grok 4.7 scores higher on the Artificial Analysis Intelligence Index: 46.4 against 44.3.
  • Grok 4.7 is newer: released 21 Sep 2026, against 12 Aug 2026 for Grok 4.6.
  • On third-party benchmarks, on the results both models have: Grok 4.6 does better for writing and sticking to the facts; Grok 4.7 for agents and tool use.

Head to head

44.3
Intelligence Index
46.4
$3.00
Price per million tokens, blendedlower is better
$3.00
500,000
Context window, tokens
500,000

Every fact side by side

Grok 4.6Grok 4.7
Price per million tokens$2.00 in · $6.00 out$2.00 in · $6.00 out
Blended (3 in : 1 out)$3.00$3.00
Cheapest host, blended$3.00 (xAI)$3.00 (Amazon Bedrock)
Cached input, per million$0.50 cached input$0.50 cached input
Intelligence Index44.346.4
Spring Prompt overall–39
Output speed, tokens a second–74
Context window500,000 tokens500,000 tokens
Released12 Aug 202621 Sep 2026 (newer)
WeightsClosedClosed
Developer based inthe USthe US

Prices are the developer's list price where we have read it, otherwise the typical host on OpenRouter; see every model's price. Intelligence and speed from Artificial Analysis.

Which is better for what

Each model's rank among the models on sale today, for each kind of work, from our own benchmarks and licensed sources. Further right is better; the better rank is in green.

Product listingsTurning a sparse product feed and photos into listings that can go live not measured 11 of 35
Decks from an analysisTurning a finished analysis into a deck you could present as it is not measured 13 of 35
Marketing planningPlanning a year of ad spend without overspending not measured 27 of 35
Agents and tool useMulti-step tasks with tools: support desks, coding agents, function calls 14 of 159 33 of 159
Professional workReal tasks from banking, consulting and law, business documents and freelance projects 16 of 51 not measured
Reasoning and knowledgeHard questions across science, maths and general knowledge 16 of 187 23 of 187
WritingWhat people prefer in blind comparisons, and judged writing quality 43 of 169 62 of 169
Sticking to the factsSummarising without inventing things, and factual answers 55 of 139 102 of 139
Speed under pressureGood decisions against a real clock (fast chess) not measured 39 of 39

From our research

Every shared result

Each model's best result on every measure both have, grouped by benchmark, with the setting that produced it. Under each measure, every model that benchmark has measured is a grey tick, best to the right, so you can see where the two sit in the field. The headline measures of each benchmark come first; the rest are under "Show all". Green marks the better value; where a source reports ranges that overlap, neither is marked.

MeasureGrok 4.6Grok 4.7
Arena (formerly LMArena) reported by Arena (formerly LMArena) · 3 results
Overall · TextArena rating, higher is better 62 · 79 of 387 measured 1,454high reasoning 1,442xhigh reasoning
Overall · Text factualityArena rating, higher is better 62 · 101 of 163 measured 1,449high reasoning 1,428xhigh reasoning
Business, management and finance · TextArena rating, higher is better 72 · 95 of 380 measured 1,447high reasoning 1,433xhigh reasoning
Artificial Analysis reported by Artificial Analysis · 3 results
Artificial Analysis Intelligence Indexindex score, higher is better 18 · 13 of 249 measured 44.3high reasoning 46.4xhigh reasoning
Humanity's Last Exam% of questions, higher is better 26 · 30 of 250 measured 44.1%xhigh reasoning 43.1%xhigh reasoning
Terminal-Bench 4.0% of tasks, higher is better 29 · 23 of 120 measured 21.2%high reasoning 25.8%xhigh reasoning
UGI Leaderboard reported by UGI Leaderboard · 3 results
Writing score · UGIscore out of 100, higher is better 34 · 63 of 230 measured 63.3 54.1
Requested-length error · UGI% off the requested word count, lower is better 147 · 158 of 230 measured 25.0% 28.0%
Style adherence · UGIscore from 0 to 1, higher is better 76 · 53 of 230 measured 0.36 0.37
Show all 19 results (10 more)
MeasureGrok 4.6Grok 4.7
Arena (formerly LMArena) reported by Arena (formerly LMArena) · 8 results
Confirmed task success · AgentIPS effect estimate, higher is better 30 · 15 of 48 measured -0.01xhigh reasoning, arena agent 0.05xhigh reasoning, arena agent
Creative writing · TextArena rating, higher is better 43 · 53 of 385 measured 1,445high reasoning 1,436xhigh reasoning
Expert prompts · TextArena rating, higher is better 52 · 92 of 338 measured 1,491high reasoning 1,457xhigh reasoning
Instruction following · TextArena rating, higher is better 53 · 68 of 387 measured 1,451high reasoning 1,440xhigh reasoning
Praise over complaint · AgentIPS effect estimate, higher is better 26 · 15 of 48 measured 0.00xhigh reasoning, arena agent 0.07xhigh reasoning, arena agent
Steerability · AgentIPS effect estimate, higher is better 29 · 30 of 48 measured 0.01xhigh reasoning, arena agent 0.00xhigh reasoning, arena agent
Tool grounding · AgentIPS effect estimate, higher is better 20 · 23 of 48 measured 0.00xhigh reasoning, arena agent 0.00xhigh reasoning, arena agent
Writing, literature and language · TextArena rating, higher is better 52 · 68 of 386 measured 1,445high reasoning 1,430xhigh reasoning
Artificial Analysis reported by Artificial Analysis · 2 results
Long-context reasoning (AA-LCR)% of questions, higher is better 34 · 42 of 241 measured 81.0%medium reasoning 80.0%low reasoning
SciCode% of problems, higher is better 21 · 16 of 121 measured 56.5%high reasoning 57.8%high reasoning

Quick answers

Is Grok 4.6 better than Grok 4.7?

Grok 4.7 is the newer version at about the same price, and it scores higher (Intelligence Index 46.4 against 44.3): upgrade.

Which is cheaper, Grok 4.6 or Grok 4.7?

They cost about the same: $3.00 and $3.00 per million tokens (three input to one output).

Which is better for agents and tool use, Grok 4.6 or Grok 4.7?

Grok 4.6 ranks 14 of 159 models on sale for multi-step tasks with tools: support desks, coding agents, function calls, against 33 for Grok 4.7.

Which is better for reasoning and knowledge, Grok 4.6 or Grok 4.7?

Grok 4.6 ranks 16 of 187 models on sale for hard questions across science, maths and general knowledge, against 23 for Grok 4.7.

Which is better for writing, Grok 4.6 or Grok 4.7?

Grok 4.6 ranks 43 of 169 models on sale for what people prefer in blind comparisons, and judged writing quality, against 62 for Grok 4.7.

Which is better for sticking to the facts, Grok 4.6 or Grok 4.7?

Grok 4.6 ranks 55 of 139 models on sale for summarising without inventing things, and factual answers, against 102 for Grok 4.7.

Run Grok 4.6 and Grok 4.7 on your own prompt.

Benchmarks aren't your task. Test both side by side on your prompt, with every answer checked against your rules and the cost of each one.