Enter opens the highlighted result. Esc closes.

Grok 4.5 vs Qwen3 32B

Grok 4.5 leads Qwen3 32B on the Artificial Analysis Intelligence Index, 38.8 to 7.3, and Qwen3 32B costs 15.0x less per output token ($0.40 vs $6.00 per million). Updated October 22, 2026; 34 benchmarks have a result for both.

Replace Grok 4.5 Replace Qwen3 32B
Grok 4.5xAI
Coding
58.6
Overall
62.1
Output $/M
$6.00
34
Qwen3 32BQwen
Coding
No data
Overall
40.8
Output $/M
$0.40
0
vs boards won of 34 shared

Grok 4.5 or Qwen3 32B: which to pick

Only Grok 4.5 has a coding score, 58.6 on the Vibe Coding Index; Qwen3 32B has none, so the two cannot be ranked on coding.

Pick Qwen3 32B for output tokens 15.0x cheaper ($0.40 vs $6.00 per million).

Model AGrok 4.5xAI58.6/1006 independent benchmarks$0.62 per solved taskModel BQwen3 32BQwenNot benchmarkedNo coding benchmarkNo cost per solved task
Shared results
34
Grok 4.5 only
89
Qwen3 32B only
11
Like-for-like categories
6 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 4.5 58.6 from 6 benchmarks, Qwen3 32B not benchmarked.
Widest category
General, 26.3 points apart on average over 15 shared benchmarks.
Widest single gap
Long context reasoning (AA-LCR): Grok 4.5 79.3% against Qwen3 32B 0.0%, 79.3 points apart on a 0 to 100 scale.
Output price
Grok 4.5 $6.00 against Qwen3 32B $0.40 per million output tokens, 15.0x apart.
Context window
Grok 4.5 takes 500K tokens, Qwen3 32B 131K.

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 4.5BarsQwen3 32BTies
Coding30
Intelligence20
Reasoning70
Math30
Knowledge40
General150
All shared boards340
CodingLike-for-likeSix of these boards feed the Vibe Coding Index30
Coding benchmarks for Grok 4.5 and Qwen3 32B
BenchmarkGrok 4.5Qwen3 32BLeader
Aider polyglotNot measured40.0%Only Qwen3 32B measured
LiveBench coding94.3%Not measuredOnly Grok 4.5 measured
LMArena WebDev1553Not measuredOnly Grok 4.5 measured
AA Coding Index72.4Not measuredOnly Grok 4.5 measured
LiveCodeBench (Vals AI)87.3%Not measuredOnly Grok 4.5 measured
SWE-bench Verified (Vals AI)86.6%Not measuredOnly Grok 4.5 measured
Terminal-Bench 4 (Vals AI)8.6%Not measuredOnly Grok 4.5 measured
Vibe Code Bench (Vals AI)69.0%Not measuredOnly Grok 4.5 measured
Code migration (Vals AI)36.6%Not measuredOnly Grok 4.5 measured
ProgramBench (Vals AI)0.0%Not measuredOnly Grok 4.5 measured
IOI (Vals AI)40.6%Not measuredOnly Grok 4.5 measured
Terminal-Bench 2.1 (Vals AI)67.8%Not measuredOnly Grok 4.5 measured
SkillsBench (Vals AI)66.0%Not measuredOnly Grok 4.5 measured
CyberBench (Vals AI)71.0%Not measuredOnly Grok 4.5 measured
LiveCodeBench (Artificial Analysis run)Not measured54.6%Only Qwen3 32B measured
SciCode55.0%36.0%Grok 4.5
Terminal-Bench Hard (Artificial Analysis run)Not measured3.0%Only Qwen3 32B measured
FrontierCode42.4%Not measuredOnly Grok 4.5 measured
WeirdML46.4%Not measuredOnly Grok 4.5 measured
ALE-Bench1309.3Not measuredOnly Grok 4.5 measured
LMArena Text: coding14741357Grok 4.5
LMArena Text: software and IT14711366Grok 4.5
LMArena WebDev: React1556Not measuredOnly Grok 4.5 measured
LMArena WebDev: HTML1567Not measuredOnly Grok 4.5 measured
LMArena WebDev: from an image1562Not measuredOnly Grok 4.5 measured
SciCode6.5%Not measuredOnly Grok 4.5 measured
Neither model has a result on the other 13 boards here
IntelligenceLike-for-likePublic boards, shown as published, kept out of the index20
Intelligence benchmarks for Grok 4.5 and Qwen3 32B
BenchmarkGrok 4.5Qwen3 32BLeader
AA Intelligence Index38.87.3Grok 4.5
Epoch Capabilities Index154.1138.4Grok 4.5
Agentic workDirectional onlyPublic boards, shown as published, kept out of the index25 boards, none shared
Agentic work benchmarks for Grok 4.5 and Qwen3 32B
BenchmarkGrok 4.5Qwen3 32BLeader
AA Agentic Index41.2Not measuredOnly Grok 4.5 measured
Design Arena: agents building web apps1218Not measuredOnly Grok 4.5 measured
Design Arena: agents building full-stack apps1249Not measuredOnly Grok 4.5 measured
Design Arena: agents building mobile apps1220Not measuredOnly Grok 4.5 measured
Design Arena: agents building Android apps1236Not measuredOnly Grok 4.5 measured
Design Arena: agents building Godot games1229Not measuredOnly Grok 4.5 measured
Design Arena: agents building games1236Not measuredOnly Grok 4.5 measured
Design Arena: HTML slides1205Not measuredOnly Grok 4.5 measured
Design Arena: PowerPoint slides1230Not measuredOnly Grok 4.5 measured
LiveBench research85.9%Not measuredOnly Grok 4.5 measured
LiveBench browser44.0%Not measuredOnly Grok 4.5 measured
LiveBench long agentic27.9%Not measuredOnly Grok 4.5 measured
SREBench (Vals AI)0.8%Not measuredOnly Grok 4.5 measured
Finance Agent (Vals AI)48.4%Not measuredOnly Grok 4.5 measured
Tax Agent Bench (Vals AI)24.9%Not measuredOnly Grok 4.5 measured
Legal Research Bench (Vals AI)38.0%Not measuredOnly Grok 4.5 measured
Excel Modeling Benchmark (Vals AI)52.9%Not measuredOnly Grok 4.5 measured
Time Horizon Index (Vals AI)6.3%Not measuredOnly Grok 4.5 measured
τ²-Bench TelecomNot measured29.8%Only Qwen3 32B measured
APEX-Agents56.2%Not measuredOnly Grok 4.5 measured
PostTrainBench23.4%Not measuredOnly Grok 4.5 measured
Vending-Bench 2$3,887Not measuredOnly Grok 4.5 measured
LMArena Search1216Not measuredOnly Grok 4.5 measured
Berkeley Function Calling Leaderboard (BFCL)Not measured48.7%Only Qwen3 32B measured
BrowseComp-PlusNot measured10.7%Only Qwen3 32B measured
Neither model has a result on the other 16 boards here
ReasoningLike-for-likePublic boards, shown as published, kept out of the index70
Reasoning benchmarks for Grok 4.5 and Qwen3 32B
BenchmarkGrok 4.5Qwen3 32BLeader
GPQA Diamond92.9%Not measuredOnly Grok 4.5 measured
LiveBench reasoning87.2%Not measuredOnly Grok 4.5 measured
MMMU Pro (Vals AI)61.8%Not measuredOnly Grok 4.5 measured
GPQA Diamond (Artificial Analysis run)93.1%66.8%Grok 4.5
Humanity's Last Exam (Artificial Analysis run)42.7%7.4%Grok 4.5
GPQA Diamond (Epoch AI run)93.4%65.7%Grok 4.5
ARC-AGI87.2%Not measuredOnly Grok 4.5 measured
ARC-AGI-252.6%Not measuredOnly Grok 4.5 measured
SimpleBench70.0%Not measuredOnly Grok 4.5 measured
Chess puzzles35.0%Not measuredOnly Grok 4.5 measured
LMArena Text: hard prompts14621334Grok 4.5
LMArena Text: expert14651362Grok 4.5
LMArena Text: hard prompts in English14641350Grok 4.5
LMArena Vision: diagrams1304Not measuredOnly Grok 4.5 measured
LMArena Vision: homework1295Not measuredOnly Grok 4.5 measured
Kagi LLM Benchmark83.5%48.7%Grok 4.5
Neither model has a result on the other 6 boards here
MathLike-for-likePublic boards, shown as published, kept out of the index30
Math benchmarks for Grok 4.5 and Qwen3 32B
BenchmarkGrok 4.5Qwen3 32BLeader
AA Math IndexNot measured73.0Only Qwen3 32B measured
AIME 2025 (Artificial Analysis run)Not measured73.0%Only Qwen3 32B measured
OTIS Mock AIME97.8%66.9%Grok 4.5
FrontierMath, tiers 1 to 357.2%Not measuredOnly Grok 4.5 measured
FrontierMath, tier 424.4%Not measuredOnly Grok 4.5 measured
ProofBench31.0%Not measuredOnly Grok 4.5 measured
LMArena Text: math14601399Grok 4.5
LMArena Text: mathematical work14821420Grok 4.5
Neither model has a result on the other 5 boards here
KnowledgeLike-for-likePublic boards, shown as published, kept out of the index40
Knowledge benchmarks for Grok 4.5 and Qwen3 32B
BenchmarkGrok 4.5Qwen3 32BLeader
MMLU Pro89.2%Not measuredOnly Grok 4.5 measured
LegalBench (Vals AI)86.0%Not measuredOnly Grok 4.5 measured
CorpFin (Vals AI)67.4%Not measuredOnly Grok 4.5 measured
TaxEval (Vals AI)71.7%Not measuredOnly Grok 4.5 measured
MortgageTax (Vals AI)61.8%Not measuredOnly Grok 4.5 measured
MedCode (Vals AI)43.3%Not measuredOnly Grok 4.5 measured
MMLU Pro (Artificial Analysis run)Not measured79.8%Only Qwen3 32B measured
SimpleQA Verified48.3%Not measuredOnly Grok 4.5 measured
LMArena Text: science14641372Grok 4.5
LMArena Text: medicine and healthcare14401368Grok 4.5
LMArena Text: legal and government14631324Grok 4.5
LMArena Text: business and finance14411318Grok 4.5
FACTS Benchmark Suite50.2%Not measuredOnly Grok 4.5 measured
Neither model has a result on the other 5 boards here
Web designDirectional onlyPublic boards, shown as published, kept out of the index9 boards, none shared
Web design benchmarks for Grok 4.5 and Qwen3 32B
BenchmarkGrok 4.5Qwen3 32BLeader
Design Arena: websites1281Not measuredOnly Grok 4.5 measured
Design Arena: UI components1280Not measuredOnly Grok 4.5 measured
Design Arena: game dev1264Not measuredOnly Grok 4.5 measured
Design Arena: data viz1266Not measuredOnly Grok 4.5 measured
Design Arena: 3D1257Not measuredOnly Grok 4.5 measured
Design Arena: all code categories1275Not measuredOnly Grok 4.5 measured
Design Arena: SVG1215Not measuredOnly Grok 4.5 measured
Design Arena: ASCII art1256Not measuredOnly Grok 4.5 measured
LiveBench design78.1%Not measuredOnly Grok 4.5 measured
GeneralLike-for-likePublic boards, shown as published, kept out of the index150
General benchmarks for Grok 4.5 and Qwen3 32B
BenchmarkGrok 4.5Qwen3 32BLeader
LiveBench data analysis73.0%Not measuredOnly Grok 4.5 measured
SAGE (Vals AI)35.0%Not measuredOnly Grok 4.5 measured
MedScribe (Vals AI)86.9%Not measuredOnly Grok 4.5 measured
Vals Multimodal Index63.4%Not measuredOnly Grok 4.5 measured
IFBenchNot measured36.3%Only Qwen3 32B measured
Long context reasoning (AA-LCR)79.3%0.0%Grok 4.5
DTBench96.5%67.5%Grok 4.5
LMCA45.2%17.3%Grok 4.5
Fiction.LiveBenchNot measured74.2%Only Qwen3 32B measured
Surface Evolver Bench74.4%Not measuredOnly Grok 4.5 measured
Furniture assembly22.5%Not measuredOnly Grok 4.5 measured
LMArena Text14481340Grok 4.5
LMArena Text: instruction following14451305Grok 4.5
LMArena Text: creative writing14421297Grok 4.5
LMArena Vision1287Not measuredOnly Grok 4.5 measured
LMArena Text: English14531359Grok 4.5
LMArena Text: non-English14401317Grok 4.5
LMArena Text: Chinese14951357Grok 4.5
LMArena Text: German14481341Grok 4.5
LMArena Text: Japanese1430Not measuredOnly Grok 4.5 measured
LMArena Text: Korean1406Not measuredOnly Grok 4.5 measured
LMArena Text: French1457Not measuredOnly Grok 4.5 measured
LMArena Text: Spanish1451Not measuredOnly Grok 4.5 measured
LMArena Text: Russian14491311Grok 4.5
LMArena Text: Polish1432Not measuredOnly Grok 4.5 measured
LMArena Text: multi-turn14561331Grok 4.5
LMArena Text: longer queries14621327Grok 4.5
LMArena Text: writing and language14451298Grok 4.5
LMArena Text: entertainment and media14341279Grok 4.5
LMArena Vision: English1285Not measuredOnly Grok 4.5 measured
LMArena Vision: Chinese1345Not measuredOnly Grok 4.5 measured
LMArena Vision: reading text1293Not measuredOnly Grok 4.5 measured
LMArena Vision: humor1276Not measuredOnly Grok 4.5 measured
LMArena Vision: creative writing1293Not measuredOnly Grok 4.5 measured
LMArena Document1452Not measuredOnly Grok 4.5 measured
Neither model has a result on the other 11 boards here
Image generationNot comparablePublic boards, shown as published, kept out of the index0 boards, none shared
Image generation benchmarks for Grok 4.5 and Qwen3 32B
BenchmarkGrok 4.5Qwen3 32BLeader
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 4.5$0.0050Fits in one request
Qwen3 32B$0.0003Fits in one request

Qwen3 32B costs less

Repository review

50K input + 3K output tokens

Grok 4.5$0.118Fits in one request
Qwen3 32B$0.0082Fits in one request

Qwen3 32B costs less

Agent loop

200K input + 10K output tokens

Grok 4.5$0.460Fits in one request
Qwen3 32B$0.032Exceeds the context window

Qwen3 32B 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 4.5 and Qwen3 32B
SpecificationSpaceXAI Grok 4.5Qwen3 32B
MakerxAIQwen
WeightsNo public weights listedOpen weights
ReleasedJuly 8, 2026April 28, 2025
Context window500K131K
Input $/Mtok$2.00$0.14
Output $/Mtok$6.00$0.40
Coding index58.6Not benchmarked
Independent measurements behind it60
Cost per solved task$0.62Not computable
SourceOpenRouter listingOpenRouter listing

Questions

Is Grok 4.5 better than Qwen3 32B for coding?

Only Grok 4.5 has a coding score, 58.6 on the Vibe Coding Index; Qwen3 32B has none, so the two cannot be ranked on coding. Pick Qwen3 32B for output tokens 15.0x cheaper ($0.40 vs $6.00 per million).

Which is cheaper, Grok 4.5 or Qwen3 32B?

Grok 4.5 costs $2.00 per million input tokens and $6.00 per million output tokens; Qwen3 32B costs $0.14 and $0.40. Cost per solved task is known for Grok 4.5 only.

Which has the bigger context window, Grok 4.5 or Qwen3 32B?

Grok 4.5, with 500K tokens against 131K.

How many benchmarks is the Grok 4.5 vs Qwen3 32B comparison based on?

34 benchmarks have a result for both models as of October 22, 2026. Grok 4.5's coding index rests on 6 independent benchmarks and Qwen3 32B's on 0. Where only one model has a result, the page says so instead of naming a winner.

Compare next