Enter opens the highlighted result. Esc closes.

Grok 4.6 vs Qwen3.8 27B

Qwen3.8 27B leads Grok 4.6 on coding, 74.2 to 65.7 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; 83 benchmarks have a result for both.

Replace Grok 4.6 Replace Qwen3.8 27B
Grok 4.6xAI
Coding
65.7
Overall
67.4
Output $/M
$6.00
52
Qwen3.8 27BQwen
Coding
74.2
Overall
59.6
Output $/M
$2.55
30
vs boards won of 83 shared, 1 tied

Grok 4.6 or Qwen3.8 27B: which to pick

Pick Qwen3.8 27B for coding: it scores 74.2 on the Vibe Coding Index against 65.7 for Grok 4.6, though on 4 benchmarks against 6.

Pick Grok 4.6 for a lower cost per solved task ($0.82 against $1.26) and work measured by LiveBench coding or LMArena WebDev, where it leads.

Model AGrok 4.6xAI65.7/1006 independent benchmarks$0.82 per solved taskModel BQwen3.8 27BQwen74.2/1004 independent benchmarks$1.26 per solved task
Shared results
83
Grok 4.6 only
40
Qwen3.8 27B only
2
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 4.6 65.7 from 6 benchmarks, Qwen3.8 27B 74.2 from 4 benchmarks.
Widest category
Math, 14.1 points apart on average over 3 shared benchmarks.
Widest single gap
ProofBench: Grok 4.6 51.0% against Qwen3.8 27B 16.0%, 35.0 points apart on a 0 to 100 scale.
Output price
Grok 4.6 $6.00 against Qwen3.8 27B $2.55 per million output tokens, 2.4x apart.
Context window
Grok 4.6 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 4.6BarsQwen3.8 27BTies
Coding153
Intelligence20
Agentic work100
Reasoning641 tie
Math21
Knowledge45
Web design10
General1217
All shared boards52301 tie
CodingLike-for-likeSix of these boards feed the Vibe Coding Index153
Coding benchmarks for Grok 4.6 and Qwen3.8 27B
BenchmarkGrok 4.6Qwen3.8 27BLeader
LiveBench coding89.1%86.8%Grok 4.6
LMArena WebDev16171593Grok 4.6
AA Coding Index76.868.1Grok 4.6
LiveCodeBench (Vals AI)88.2%84.0%Grok 4.6
SWE-bench Verified (Vals AI)95.6%86.0%Grok 4.6
Terminal-Bench 4 (Vals AI)17.2%Not measuredOnly Grok 4.6 measured
Vibe Code Bench (Vals AI)76.2%64.8%Grok 4.6
Code migration (Vals AI)44.6%14.2%Grok 4.6
ProgramBench (Vals AI)1.0%0.0%Grok 4.6
IOI (Vals AI)47.6%39.1%Grok 4.6
Terminal-Bench 2.1 (Vals AI)78.3%58.4%Grok 4.6
Vibe Code Bench 1-100 (Vals AI)14.8%Not measuredOnly Grok 4.6 measured
SkillsBench (Vals AI)55.8%38.1%Grok 4.6
CyberBench (Vals AI)67.7%67.8%Qwen3.8 27B
SciCode56.5%46.6%Grok 4.6
FrontierCode48.0%Not measuredOnly Grok 4.6 measured
FrontierSWE25.3%Not measuredOnly Grok 4.6 measured
WeirdML67.3%Not measuredOnly Grok 4.6 measured
ALE-Bench1508.1Not measuredOnly Grok 4.6 measured
LMArena Text: coding14661483Qwen3.8 27B
LMArena Text: software and IT14571478Qwen3.8 27B
LMArena WebDev: React16311602Grok 4.6
LMArena WebDev: HTML15971570Grok 4.6
LMArena WebDev: from an image16031577Grok 4.6
Neither model has a result on the other 15 boards here
IntelligenceLike-for-likePublic boards, shown as published, kept out of the index20
Intelligence benchmarks for Grok 4.6 and Qwen3.8 27B
BenchmarkGrok 4.6Qwen3.8 27BLeader
AA Intelligence Index44.333.7Grok 4.6
Epoch Capabilities Index156.6151.7Grok 4.6
Agentic workLike-for-likePublic boards, shown as published, kept out of the index100
Agentic work benchmarks for Grok 4.6 and Qwen3.8 27B
BenchmarkGrok 4.6Qwen3.8 27BLeader
AA Agentic Index53.045.8Grok 4.6
Design Arena: agents building web apps1248Not measuredOnly Grok 4.6 measured
Design Arena: agents building full-stack apps1264Not measuredOnly Grok 4.6 measured
Design Arena: agents building mobile apps1234Not measuredOnly Grok 4.6 measured
Design Arena: agents building Android apps1291Not measuredOnly Grok 4.6 measured
Design Arena: agents building Godot games1276Not measuredOnly Grok 4.6 measured
Design Arena: agents building games1209Not measuredOnly Grok 4.6 measured
Design Arena: HTML slides1250Not measuredOnly Grok 4.6 measured
Design Arena: PowerPoint slides1237Not measuredOnly Grok 4.6 measured
LiveBench research84.5%59.4%Grok 4.6
LiveBench browser64.0%32.0%Grok 4.6
LiveBench long agentic40.7%27.9%Grok 4.6
Finance Agent (Vals AI)53.7%48.5%Grok 4.6
Tax Agent Bench (Vals AI)33.1%30.9%Grok 4.6
Legal Research Bench (Vals AI)48.1%36.1%Grok 4.6
Excel Modeling Benchmark (Vals AI)62.7%59.7%Grok 4.6
BioMysteryBench (Vals AI)72.2%Not measuredOnly Grok 4.6 measured
Terminal-Bench-Science (Vals AI)11.4%0.0%Grok 4.6
Vals Science Index38.1%Not measuredOnly Grok 4.6 measured
Time Horizon Index (Vals AI)5.8%Not measuredOnly Grok 4.6 measured
APEX-Agents65.3%47.5%Grok 4.6
Vending-Bench 2$9,047Not measuredOnly Grok 4.6 measured
Neither model has a result on the other 19 boards here
ReasoningLike-for-likePublic boards, shown as published, kept out of the index64
Reasoning benchmarks for Grok 4.6 and Qwen3.8 27B
BenchmarkGrok 4.6Qwen3.8 27BLeader
GPQA Diamond94.7%88.9%Grok 4.6
LiveBench reasoning91.2%80.0%Grok 4.6
MysteryMechanism (Vals AI)30.6%Not measuredOnly Grok 4.6 measured
MMMU Pro (Vals AI)Not measured83.9%Only Qwen3.8 27B measured
GPQA Diamond (Artificial Analysis run)94.9%90.5%Grok 4.6
Humanity's Last Exam (Artificial Analysis run)44.1%33.9%Grok 4.6
GPQA Diamond (Epoch AI run)94.0%Not measuredOnly Grok 4.6 measured
ARC-AGI87.5%87.5%Tie
ARC-AGI-267.1%42.4%Grok 4.6
SimpleBench75.9%Not measuredOnly Grok 4.6 measured
Chess puzzles33.0%Not measuredOnly Grok 4.6 measured
Mystery game puzzles34.0%Not measuredOnly Grok 4.6 measured
EBR-bench30.5%Not measuredOnly Grok 4.6 measured
LMArena Text: hard prompts14481459Qwen3.8 27B
LMArena Text: expert14691479Qwen3.8 27B
LMArena Text: hard prompts in English14451461Qwen3.8 27B
LMArena Vision: diagrams12851285Grok 4.6
LMArena Vision: homework12831295Qwen3.8 27B
Neither model has a result on the other 4 boards here
MathLike-for-likePublic boards, shown as published, kept out of the index21
Math benchmarks for Grok 4.6 and Qwen3.8 27B
BenchmarkGrok 4.6Qwen3.8 27BLeader
OTIS Mock AIME99.2%Not measuredOnly Grok 4.6 measured
FrontierMath, tiers 1 to 366.0%Not measuredOnly Grok 4.6 measured
FrontierMath, tier 431.7%Not measuredOnly Grok 4.6 measured
ProofBench51.0%16.0%Grok 4.6
LMArena Text: math14171455Qwen3.8 27B
LMArena Text: mathematical work14641458Grok 4.6
Neither model has a result on the other 7 boards here
KnowledgeLike-for-likePublic boards, shown as published, kept out of the index45
Knowledge benchmarks for Grok 4.6 and Qwen3.8 27B
BenchmarkGrok 4.6Qwen3.8 27BLeader
MMLU Pro89.4%84.3%Grok 4.6
LegalBench (Vals AI)86.3%82.4%Grok 4.6
CorpFin (Vals AI)66.2%Not measuredOnly Grok 4.6 measured
TaxEval (Vals AI)71.1%70.8%Grok 4.6
MortgageTax (Vals AI)64.2%64.9%Qwen3.8 27B
MedCode (Vals AI)44.7%28.7%Grok 4.6
SimpleQA Verified49.3%Not measuredOnly Grok 4.6 measured
LMArena Text: science14371467Qwen3.8 27B
LMArena Text: medicine and healthcare14201437Qwen3.8 27B
LMArena Text: legal and government14291447Qwen3.8 27B
LMArena Text: business and finance14161438Qwen3.8 27B
Neither model has a result on the other 7 boards here
Web designLike-for-likePublic boards, shown as published, kept out of the index10
Web design benchmarks for Grok 4.6 and Qwen3.8 27B
BenchmarkGrok 4.6Qwen3.8 27BLeader
Design Arena: websites1289Not measuredOnly Grok 4.6 measured
Design Arena: UI components1294Not measuredOnly Grok 4.6 measured
Design Arena: game dev1296Not measuredOnly Grok 4.6 measured
Design Arena: data viz1282Not measuredOnly Grok 4.6 measured
Design Arena: 3D1277Not measuredOnly Grok 4.6 measured
Design Arena: all code categories1292Not measuredOnly Grok 4.6 measured
Design Arena: SVG1240Not measuredOnly Grok 4.6 measured
Design Arena: ASCII art1279Not measuredOnly Grok 4.6 measured
LiveBench design79.8%75.9%Grok 4.6
GeneralLike-for-likePublic boards, shown as published, kept out of the index1217
General benchmarks for Grok 4.6 and Qwen3.8 27B
BenchmarkGrok 4.6Qwen3.8 27BLeader
LiveBench data analysis75.3%76.6%Qwen3.8 27B
Public Benefits Bench (Vals AI)66.8%Not measuredOnly Grok 4.6 measured
SAGE (Vals AI)28.9%52.4%Qwen3.8 27B
MedScribe (Vals AI)86.5%83.8%Grok 4.6
Long context reasoning (AA-LCR)81.0%82.0%Qwen3.8 27B
DTBench97.3%88.0%Grok 4.6
LMCA48.5%41.4%Grok 4.6
Surface Evolver BenchNot measured45.0%Only Qwen3.8 27B measured
Furniture assembly40.0%Not measuredOnly Grok 4.6 measured
LMArena Text14291440Qwen3.8 27B
LMArena Text: instruction following14311436Qwen3.8 27B
LMArena Text: creative writing14261383Grok 4.6
LMArena Vision12631269Qwen3.8 27B
LMArena Text: English14331448Qwen3.8 27B
LMArena Text: non-English14201428Qwen3.8 27B
LMArena Text: Chinese14761501Qwen3.8 27B
LMArena Text: German14261443Qwen3.8 27B
LMArena Text: Japanese13781387Qwen3.8 27B
LMArena Text: Korean13911384Grok 4.6
LMArena Text: French14551465Qwen3.8 27B
LMArena Text: Spanish14141449Qwen3.8 27B
LMArena Text: Russian14211415Grok 4.6
LMArena Text: Polish14301439Qwen3.8 27B
LMArena Text: multi-turn14251440Qwen3.8 27B
LMArena Text: longer queries14561448Grok 4.6
LMArena Text: writing and language14291405Grok 4.6
LMArena Text: entertainment and media14211393Grok 4.6
LMArena Vision: English12701267Grok 4.6
LMArena Vision: Chinese12891326Qwen3.8 27B
LMArena Vision: reading text12721281Qwen3.8 27B
LMArena Vision: humor12701235Grok 4.6
LMArena Vision: creative writing12621237Grok 4.6
LMArena Document1452Not measuredOnly Grok 4.6 measured
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 4.6 and Qwen3.8 27B
BenchmarkGrok 4.6Qwen3.8 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 4.6$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 4.6$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 4.6$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 4.6 and Qwen3.8 27B
SpecificationSpaceXAI Grok 4.6Qwen3.8 27B
MakerxAIQwen
WeightsNo public weights listedOpen weights
ReleasedAugust 12, 2026August 14, 2026
Context window500K1M
Input $/Mtok$2.00$0.425
Output $/Mtok$6.00$2.55
Coding index65.774.2
Independent measurements behind it64
Cost per solved task$0.82$1.26
SourceOpenRouter listingOpenRouter listing

Questions

Is Grok 4.6 better than Qwen3.8 27B for coding?

Pick Qwen3.8 27B for coding: it scores 74.2 on the Vibe Coding Index against 65.7 for Grok 4.6, though on 4 benchmarks against 6. Pick Grok 4.6 for a lower cost per solved task ($0.82 against $1.26) and work measured by LiveBench coding or LMArena WebDev, where it leads.

Which is cheaper, Grok 4.6 or Qwen3.8 27B?

Grok 4.6 costs $2.00 per million input tokens and $6.00 per million output tokens; Qwen3.8 27B costs $0.425 and $2.55. Per solved SWE-bench Verified task, Grok 4.6 costs $0.82 and Qwen3.8 27B $1.26.

Which has the bigger context window, Grok 4.6 or Qwen3.8 27B?

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

How many benchmarks is the Grok 4.6 vs Qwen3.8 27B comparison based on?

83 benchmarks have a result for both models as of October 22, 2026. Grok 4.6's coding index rests on 6 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-4-6-vs-qwen3-8-27b" title="Grok 4.6 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-4-6-vs-qwen3-8-27b">LLMBenchmark.ai, Grok 4.6 vs Qwen3.8 27B</a></p>