Enter opens the highlighted result. Esc closes.

Grok 4.7 vs Qwen3.8 27B

Qwen3.8 27B leads Grok 4.7 on coding, 74.2 to 66.7 on the Vibe Coding Index, and Qwen3.8 27B costs 2.4x less per output token ($2.55 vs $6.00 per million). Updated October 22, 2026; 72 benchmarks have a result for both.

Replace Grok 4.7 Replace Qwen3.8 27B
Grok 4.7xAI
Coding
66.7
Overall
66.6
Output $/M
$6.00
35
Qwen3.8 27BQwen
Coding
74.2
Overall
59.6
Output $/M
$2.55
37
vs boards won of 72 shared

Grok 4.7 or Qwen3.8 27B: which to pick

Pick Qwen3.8 27B for coding: it scores 74.2 on the Vibe Coding Index against 66.7 for Grok 4.7, though on 4 benchmarks against 5.

Pick Grok 4.7 for work measured by LMArena WebDev or Vibe Code Bench (Vals AI), where it leads.

Model AGrok 4.7xAI66.7/1005 independent benchmarksNo cost per solved taskModel BQwen3.8 27BQwen74.2/1004 independent benchmarks$1.26 per solved task
Shared results
72
Grok 4.7 only
27
Qwen3.8 27B only
13
Like-for-like categories
8 of 9

Only benchmarks that tested both models count toward a pick; a result only one model has is shown but not compared.

Which one for your work

Each card is decided by one number both models carry. Where only one has it, the card says so and makes no call.

At a glance

Coding index
Grok 4.7 66.7 from 5 benchmarks, Qwen3.8 27B 74.2 from 4 benchmarks.
Widest category
Agentic work, 13.3 points apart on average over 9 shared benchmarks.
Widest single gap
Code migration (Vals AI): Grok 4.7 44.8% against Qwen3.8 27B 14.2%, 30.7 points apart on a 0 to 100 scale.
Output price
Grok 4.7 $6.00 against Qwen3.8 27B $2.55 per million output tokens, 2.4x apart.
Context window
Grok 4.7 takes 500K tokens, Qwen3.8 27B 1M.

Benchmark by benchmark

First the score: boards each model leads, per category, where both have a result. Then every board, a category to a fold; bars are each result on a 0 to 100 scale, so compare the two in a row, not one row with another.

Boards won, where both have a result
CategoryGrok 4.7BarsQwen3.8 27BTies
Coding113
Intelligence20
Agentic work90
Reasoning54
Math12
Knowledge24
Web design10
General424
All shared boards3537
CodingLike-for-likeSix of these boards feed the Vibe Coding Index113
Coding benchmarks for Grok 4.7 and Qwen3.8 27B
BenchmarkGrok 4.7Qwen3.8 27BLeader
LiveBench coding84.5%86.8%Qwen3.8 27B
LMArena WebDev16371593Grok 4.7
AA Coding IndexNot measured68.1Only Qwen3.8 27B measured
LiveCodeBench (Vals AI)Not measured84.0%Only Qwen3.8 27B measured
SWE-bench Verified (Vals AI)Not measured86.0%Only Qwen3.8 27B measured
Terminal-Bench 4 (Vals AI)28.8%Not measuredOnly Grok 4.7 measured
Vibe Code Bench (Vals AI)86.2%64.8%Grok 4.7
Code migration (Vals AI)44.8%14.2%Grok 4.7
ProgramBench (Vals AI)0.5%0.0%Grok 4.7
IOI (Vals AI)57.7%39.1%Grok 4.7
Terminal-Bench 2.1 (Vals AI)73.4%58.4%Grok 4.7
SWE-bench Multimodal (Vals AI)42.9%Not measuredOnly Grok 4.7 measured
SkillsBench (Vals AI)Not measured38.1%Only Qwen3.8 27B measured
CyberBench (Vals AI)69.5%67.8%Grok 4.7
SciCode57.8%46.6%Grok 4.7
FrontierCode47.6%Not measuredOnly Grok 4.7 measured
FrontierSWE29.5%Not measuredOnly Grok 4.7 measured
LMArena Text: coding14321483Qwen3.8 27B
LMArena Text: software and IT14211478Qwen3.8 27B
LMArena WebDev: React16631602Grok 4.7
LMArena WebDev: HTML16271570Grok 4.7
LMArena WebDev: from an image15861577Grok 4.7
Neither model has a result on the other 17 boards here
IntelligenceLike-for-likePublic boards, shown as published, kept out of the index20
Intelligence benchmarks for Grok 4.7 and Qwen3.8 27B
BenchmarkGrok 4.7Qwen3.8 27BLeader
AA Intelligence Index46.433.7Grok 4.7
Epoch Capabilities Index154.4151.7Grok 4.7
Agentic workLike-for-likePublic boards, shown as published, kept out of the index90
Agentic work benchmarks for Grok 4.7 and Qwen3.8 27B
BenchmarkGrok 4.7Qwen3.8 27BLeader
AA Agentic IndexNot measured45.8Only Qwen3.8 27B measured
LiveBench research86.4%59.4%Grok 4.7
LiveBench browser60.0%32.0%Grok 4.7
LiveBench long agentic50.1%27.9%Grok 4.7
SREBench (Vals AI)3.8%Not measuredOnly Grok 4.7 measured
Finance Agent (Vals AI)52.3%48.5%Grok 4.7
Tax Agent Bench (Vals AI)34.2%30.9%Grok 4.7
Legal Research Bench (Vals AI)47.1%36.1%Grok 4.7
Harvey's Legal Agent Benchmark (Vals AI)18.3%Not measuredOnly Grok 4.7 measured
Excel Modeling Benchmark (Vals AI)67.0%59.7%Grok 4.7
BioMysteryBench (Vals AI)69.3%Not measuredOnly Grok 4.7 measured
Terminal-Bench-Science (Vals AI)10.0%0.0%Grok 4.7
Vals Science Index34.8%Not measuredOnly Grok 4.7 measured
Robo-bench (Vals AI)0.2%Not measuredOnly Grok 4.7 measured
APEX-Agents54.6%47.5%Grok 4.7
Vending-Bench 2$10,537Not measuredOnly Grok 4.7 measured
Neither model has a result on the other 25 boards here
ReasoningLike-for-likePublic boards, shown as published, kept out of the index54
Reasoning benchmarks for Grok 4.7 and Qwen3.8 27B
BenchmarkGrok 4.7Qwen3.8 27BLeader
GPQA DiamondNot measured88.9%Only Qwen3.8 27B measured
LiveBench reasoning82.7%80.0%Grok 4.7
MysteryMechanism (Vals AI)25.2%Not measuredOnly Grok 4.7 measured
MMMU Pro (Vals AI)Not measured83.9%Only Qwen3.8 27B measured
GPQA Diamond (Artificial Analysis run)Not measured90.5%Only Qwen3.8 27B measured
Humanity's Last Exam (Artificial Analysis run)43.1%33.9%Grok 4.7
GPQA Diamond (Epoch AI run)92.7%Not measuredOnly Grok 4.7 measured
ARC-AGI90.2%87.5%Grok 4.7
ARC-AGI-261.4%42.4%Grok 4.7
Chess puzzles41.0%Not measuredOnly Grok 4.7 measured
Mystery game puzzles29.0%Not measuredOnly Grok 4.7 measured
LMArena Text: hard prompts14151459Qwen3.8 27B
LMArena Text: expert14391479Qwen3.8 27B
LMArena Text: hard prompts in English14271461Qwen3.8 27B
LMArena Vision: diagrams12541285Qwen3.8 27B
LMArena Vision: homework13091295Grok 4.7
Neither model has a result on the other 6 boards here
MathLike-for-likePublic boards, shown as published, kept out of the index12
Math benchmarks for Grok 4.7 and Qwen3.8 27B
BenchmarkGrok 4.7Qwen3.8 27BLeader
OTIS Mock AIME98.1%Not measuredOnly Grok 4.7 measured
FrontierMath, tiers 1 to 353.0%Not measuredOnly Grok 4.7 measured
FrontierMath, tier 417.1%Not measuredOnly Grok 4.7 measured
ProofBench26.0%16.0%Grok 4.7
LMArena Text: math14141455Qwen3.8 27B
LMArena Text: mathematical work14351458Qwen3.8 27B
Neither model has a result on the other 7 boards here
KnowledgeLike-for-likePublic boards, shown as published, kept out of the index24
Knowledge benchmarks for Grok 4.7 and Qwen3.8 27B
BenchmarkGrok 4.7Qwen3.8 27BLeader
MMLU ProNot measured84.3%Only Qwen3.8 27B measured
LegalBench (Vals AI)84.4%82.4%Grok 4.7
TaxEval (Vals AI)Not measured70.8%Only Qwen3.8 27B measured
MortgageTax (Vals AI)Not measured64.9%Only Qwen3.8 27B measured
MedCode (Vals AI)49.5%28.7%Grok 4.7
SimpleQA Verified56.0%Not measuredOnly Grok 4.7 measured
LMArena Text: science14191467Qwen3.8 27B
LMArena Text: medicine and healthcare13951437Qwen3.8 27B
LMArena Text: legal and government14231447Qwen3.8 27B
LMArena Text: business and finance13891438Qwen3.8 27B
Neither model has a result on the other 8 boards here
Web designLike-for-likePublic boards, shown as published, kept out of the index10
Web design benchmarks for Grok 4.7 and Qwen3.8 27B
BenchmarkGrok 4.7Qwen3.8 27BLeader
Design Arena: websites1255Not measuredOnly Grok 4.7 measured
Design Arena: UI components1275Not measuredOnly Grok 4.7 measured
Design Arena: game dev1314Not measuredOnly Grok 4.7 measured
Design Arena: data viz1238Not measuredOnly Grok 4.7 measured
Design Arena: 3D1274Not measuredOnly Grok 4.7 measured
Design Arena: all code categories1273Not measuredOnly Grok 4.7 measured
LiveBench design78.0%75.9%Grok 4.7
Neither model has a result on the other 2 boards here
GeneralLike-for-likePublic boards, shown as published, kept out of the index424
General benchmarks for Grok 4.7 and Qwen3.8 27B
BenchmarkGrok 4.7Qwen3.8 27BLeader
Vals Index58.6%Not measuredOnly Grok 4.7 measured
LiveBench data analysis76.9%76.6%Grok 4.7
Public Benefits Bench (Vals AI)65.6%Not measuredOnly Grok 4.7 measured
SAGE (Vals AI)40.8%52.4%Qwen3.8 27B
MedScribe (Vals AI)89.4%83.8%Grok 4.7
Long context reasoning (AA-LCR)80.0%82.0%Qwen3.8 27B
DTBench96.0%88.0%Grok 4.7
LMCA49.4%41.4%Grok 4.7
Surface Evolver BenchNot measured45.0%Only Qwen3.8 27B measured
Furniture assembly20.8%Not measuredOnly Grok 4.7 measured
LMArena Text13991440Qwen3.8 27B
LMArena Text: instruction following14061436Qwen3.8 27B
LMArena Text: creative writing13811383Qwen3.8 27B
LMArena Vision12311269Qwen3.8 27B
LMArena Text: English14111448Qwen3.8 27B
LMArena Text: non-English13861428Qwen3.8 27B
LMArena Text: Chinese14491501Qwen3.8 27B
LMArena Text: German13661443Qwen3.8 27B
LMArena Text: Japanese13591387Qwen3.8 27B
LMArena Text: Korean13831384Qwen3.8 27B
LMArena Text: French14311465Qwen3.8 27B
LMArena Text: Spanish14061449Qwen3.8 27B
LMArena Text: Russian13991415Qwen3.8 27B
LMArena Text: Polish14201439Qwen3.8 27B
LMArena Text: multi-turn13941440Qwen3.8 27B
LMArena Text: longer queries14161448Qwen3.8 27B
LMArena Text: writing and language13941405Qwen3.8 27B
LMArena Text: entertainment and media13761393Qwen3.8 27B
LMArena Vision: English12291267Qwen3.8 27B
LMArena Vision: Chinese12501326Qwen3.8 27B
LMArena Vision: reading text12411281Qwen3.8 27B
LMArena Vision: humorNot measured1235Only Qwen3.8 27B measured
LMArena Vision: creative writing12171237Qwen3.8 27B
Neither model has a result on the other 13 boards here
Image generationNot comparablePublic boards, shown as published, kept out of the index0 boards, none shared
Image generation benchmarks for Grok 4.7 and Qwen3.8 27B
BenchmarkGrok 4.7Qwen3.8 27BLeader
Neither model has a result on 5 boards here

What each workload costs

Three fixed token mixes at each model's listed price, so per-token rates become a bill.

Chat turn

1K input + 500 output tokens

Grok 4.7$0.0050Fits in one request
Qwen3.8 27B$0.0017Fits in one request

Qwen3.8 27B costs less

Repository review

50K input + 3K output tokens

Grok 4.7$0.118Fits in one request
Qwen3.8 27B$0.029Fits in one request

Qwen3.8 27B costs less

Agent loop

200K input + 10K output tokens

Grok 4.7$0.460Fits in one request
Qwen3.8 27B$0.111Fits in one request

Qwen3.8 27B costs less

Listed OpenRouter rates, without cache or batch discounts. A workload that does not fit a model's context window would need splitting.

Specification differences

The leading value in a row is marked; a row with a missing value names no leader. Our two rows are marked in red.

Specifications of Grok 4.7 and Qwen3.8 27B
SpecificationSpaceXAI Grok 4.7Qwen3.8 27B
MakerxAIQwen
WeightsNo public weights listedOpen weights
ReleasedSeptember 21, 2026August 14, 2026
Context window500K1M
Input $/Mtok$2.00$0.425
Output $/Mtok$6.00$2.55
Coding index66.774.2
Independent measurements behind it54
Cost per solved taskNot computable$1.26
SourceOpenRouter listingOpenRouter listing

Questions

Is Grok 4.7 better than Qwen3.8 27B for coding?

Pick Qwen3.8 27B for coding: it scores 74.2 on the Vibe Coding Index against 66.7 for Grok 4.7, though on 4 benchmarks against 5. Pick Grok 4.7 for work measured by LMArena WebDev or Vibe Code Bench (Vals AI), where it leads.

Which is cheaper, Grok 4.7 or Qwen3.8 27B?

Grok 4.7 costs $2.00 per million input tokens and $6.00 per million output tokens; Qwen3.8 27B costs $0.425 and $2.55. Cost per solved task is known for Qwen3.8 27B only.

Which has the bigger context window, Grok 4.7 or Qwen3.8 27B?

Qwen3.8 27B, with 1M tokens against 500K.

How many benchmarks is the Grok 4.7 vs Qwen3.8 27B comparison based on?

72 benchmarks have a result for both models as of October 22, 2026. Grok 4.7's coding index rests on 5 independent benchmarks and Qwen3.8 27B's on 4. Where only one model has a result, the page says so instead of naming a winner.

Compare next

Embed this comparison on your site

Free under CC BY 4.0, while the source link stays visible. Preview the frame · Terms and other embeds

<iframe src="https://llmbenchmark.ai/embed/compare/grok-4-7-vs-qwen3-8-27b" title="Grok 4.7 vs Qwen3.8 27B, LLMBenchmark.ai" width="100%" height="360" style="border:0;max-width:100%" loading="lazy"></iframe>
<p style="font:13px/1.4 sans-serif;margin:6px 0 0">Source: <a href="https://llmbenchmark.ai/compare/grok-4-7-vs-qwen3-8-27b">LLMBenchmark.ai, Grok 4.7 vs Qwen3.8 27B</a></p>