Enter opens the highlighted result. Esc closes.

Llama 4 Maverick vs Qwen2.5 72B Instruct

Llama 4 Maverick leads Qwen2.5 72B Instruct on the Artificial Analysis Intelligence Index, 9.3 to 7.7, and Qwen2.5 72B Instruct costs 2.5x less per output token ($0.40 vs $1.00 per million). Updated October 22, 2026; 41 benchmarks have a result for both.

Replace Llama 4 Maverick Replace Qwen2.5 72B Instruct
Llama 4 MaverickMeta Llama
Coding
No data
Overall
21.8
Output $/M
$1.00
36
Qwen2.5 72B InstructQwen
Coding
No data
Overall
26.5
Output $/M
$0.40
5
vs boards won of 41 shared

Llama 4 Maverick or Qwen2.5 72B Instruct: which to pick

Neither model has a coding score yet, so the two cannot be ranked on coding.

Pick Llama 4 Maverick for a larger context window (524K against 33K tokens) and work measured by τ²-Bench Telecom or LMArena Text: legal and government, where it leads.

Model ALlama 4 MaverickMeta LlamaNot benchmarkedNo coding benchmark$1.491 per solved taskModel BQwen2.5 72B InstructQwenNot benchmarkedNo coding benchmarkNo cost per solved task
Shared results
41
Llama 4 Maverick only
44
Qwen2.5 72B Instruct only
12
Like-for-like categories
7 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
Llama 4 Maverick not benchmarked, Qwen2.5 72B Instruct not benchmarked.
Widest category
Agentic work, 16.7 points apart on average over 1 shared benchmark.
Widest single gap
GPQA Diamond (Artificial Analysis run): Llama 4 Maverick 67.1% against Qwen2.5 72B Instruct 49.1%, 18.0 points apart on a 0 to 100 scale.
Output price
Llama 4 Maverick $1.00 against Qwen2.5 72B Instruct $0.40 per million output tokens, 2.5x apart.
Context window
Llama 4 Maverick takes 524K tokens, Qwen2.5 72B Instruct 33K.

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
CategoryLlama 4 MaverickBarsQwen2.5 72B InstructTies
Coding50
Intelligence20
Agentic work01
Reasoning50
Math40
Knowledge41
General163
All shared boards365
CodingLike-for-likeSix of these boards feed the Vibe Coding Index50
Coding benchmarks for Llama 4 Maverick and Qwen2.5 72B Instruct
BenchmarkLlama 4 MaverickQwen2.5 72B InstructLeader
Aider polyglot15.6%Not measuredOnly Llama 4 Maverick measured
SWE-bench Verified21.0%Not measuredOnly Llama 4 Maverick measured
AA Coding Index16.3Not measuredOnly Llama 4 Maverick measured
LiveCodeBench (Vals AI)47.3%Not measuredOnly Llama 4 Maverick measured
LiveCodeBench (Artificial Analysis run)39.7%27.6%Llama 4 Maverick
SciCode31.7%Not measuredOnly Llama 4 Maverick measured
Terminal-Bench Hard (Artificial Analysis run)6.8%4.5%Llama 4 Maverick
WeirdML24.5%16.0%Llama 4 Maverick
ALE-Bench173.0Not measuredOnly Llama 4 Maverick measured
LMArena Text: coding13011292Llama 4 Maverick
LMArena Text: software and IT12991292Llama 4 Maverick
Neither model has a result on the other 28 boards here
IntelligenceLike-for-likePublic boards, shown as published, kept out of the index20
Intelligence benchmarks for Llama 4 Maverick and Qwen2.5 72B Instruct
BenchmarkLlama 4 MaverickQwen2.5 72B InstructLeader
AA Intelligence Index9.37.7Llama 4 Maverick
Epoch Capabilities Index132.2129.0Llama 4 Maverick
Agentic workLike-for-likePublic boards, shown as published, kept out of the index01
Agentic work benchmarks for Llama 4 Maverick and Qwen2.5 72B Instruct
BenchmarkLlama 4 MaverickQwen2.5 72B InstructLeader
AA Agentic Index0.6Not measuredOnly Llama 4 Maverick measured
τ²-Bench Telecom17.8%34.5%Qwen2.5 72B Instruct
METR time horizonsNot measured35.8%Only Qwen2.5 72B Instruct measured
BALROGNot measured16.2%Only Qwen2.5 72B Instruct measured
TheAgentCompanyNot measured5.7%Only Qwen2.5 72B Instruct measured
Neither model has a result on the other 36 boards here
ReasoningLike-for-likePublic boards, shown as published, kept out of the index50
Reasoning benchmarks for Llama 4 Maverick and Qwen2.5 72B Instruct
BenchmarkLlama 4 MaverickQwen2.5 72B InstructLeader
GPQA Diamond69.4%Not measuredOnly Llama 4 Maverick measured
MMMU Pro (Vals AI)71.7%Not measuredOnly Llama 4 Maverick measured
GPQA Diamond (Artificial Analysis run)67.1%49.1%Llama 4 Maverick
Humanity's Last Exam (Artificial Analysis run)4.9%3.6%Llama 4 Maverick
GPQA Diamond (Epoch AI run)Not measured49.1%Only Qwen2.5 72B Instruct measured
ARC-AGI4.4%Not measuredOnly Llama 4 Maverick measured
ARC-AGI-20.0%Not measuredOnly Llama 4 Maverick measured
SimpleBench27.7%Not measuredOnly Llama 4 Maverick measured
BIG-Bench HardNot measured79.8%Only Qwen2.5 72B Instruct measured
LMArena Text: hard prompts12811271Llama 4 Maverick
LMArena Text: expert12591245Llama 4 Maverick
LMArena Text: hard prompts in English12951279Llama 4 Maverick
LMArena Vision: diagrams1134Not measuredOnly Llama 4 Maverick measured
LMArena Vision: homework1164Not measuredOnly Llama 4 Maverick measured
Kagi LLM Benchmark55.9%Not measuredOnly Llama 4 Maverick measured
Neither model has a result on the other 7 boards here
MathLike-for-likePublic boards, shown as published, kept out of the index40
Math benchmarks for Llama 4 Maverick and Qwen2.5 72B Instruct
BenchmarkLlama 4 MaverickQwen2.5 72B InstructLeader
AIME25.2%Not measuredOnly Llama 4 Maverick measured
MATH 500 (Vals AI)85.2%Not measuredOnly Llama 4 Maverick measured
MGSM (Vals AI)92.4%Not measuredOnly Llama 4 Maverick measured
AA Math Index19.314.0Llama 4 Maverick
AIME 2025 (Artificial Analysis run)19.3%14.0%Llama 4 Maverick
OTIS Mock AIMENot measured8.1%Only Qwen2.5 72B Instruct measured
MATH level 5Not measured63.2%Only Qwen2.5 72B Instruct measured
LMArena Text: math12991283Llama 4 Maverick
LMArena Text: mathematical work12951290Llama 4 Maverick
Neither model has a result on the other 4 boards here
KnowledgeLike-for-likePublic boards, shown as published, kept out of the index41
Knowledge benchmarks for Llama 4 Maverick and Qwen2.5 72B Instruct
BenchmarkLlama 4 MaverickQwen2.5 72B InstructLeader
MMLU Pro79.4%Not measuredOnly Llama 4 Maverick measured
LegalBench (Vals AI)77.8%Not measuredOnly Llama 4 Maverick measured
CorpFin (Vals AI)49.7%Not measuredOnly Llama 4 Maverick measured
TaxEval (Vals AI)66.6%Not measuredOnly Llama 4 Maverick measured
MortgageTax (Vals AI)58.5%Not measuredOnly Llama 4 Maverick measured
MedQA (Vals AI)43.3%Not measuredOnly Llama 4 Maverick measured
MedCode (Vals AI)36.5%Not measuredOnly Llama 4 Maverick measured
MMLU Pro (Artificial Analysis run)80.9%72.0%Llama 4 Maverick
MMLUNot measured85.3%Only Qwen2.5 72B Instruct measured
GeoBench52.0%Not measuredOnly Llama 4 Maverick measured
AI2 Reasoning ChallengeNot measured94.5%Only Qwen2.5 72B Instruct measured
LMArena Text: science12881275Llama 4 Maverick
LMArena Text: medicine and healthcare12751258Llama 4 Maverick
LMArena Text: legal and government12861288Qwen2.5 72B Instruct
LMArena Text: business and finance12601253Llama 4 Maverick
Neither model has a result on the other 3 boards here
Web designDirectional onlyPublic boards, shown as published, kept out of the index6 boards, none shared
Web design benchmarks for Llama 4 Maverick and Qwen2.5 72B Instruct
BenchmarkLlama 4 MaverickQwen2.5 72B InstructLeader
Design Arena: websites871Not measuredOnly Llama 4 Maverick measured
Design Arena: UI components896Not measuredOnly Llama 4 Maverick measured
Design Arena: game dev837Not measuredOnly Llama 4 Maverick measured
Design Arena: data viz882Not measuredOnly Llama 4 Maverick measured
Design Arena: 3D900Not measuredOnly Llama 4 Maverick measured
Design Arena: all code categories879Not measuredOnly Llama 4 Maverick measured
Neither model has a result on the other 3 boards here
GeneralLike-for-likePublic boards, shown as published, kept out of the index163
General benchmarks for Llama 4 Maverick and Qwen2.5 72B Instruct
BenchmarkLlama 4 MaverickQwen2.5 72B InstructLeader
SAGE (Vals AI)35.7%Not measuredOnly Llama 4 Maverick measured
MedScribe (Vals AI)54.2%Not measuredOnly Llama 4 Maverick measured
IFBench43.0%36.9%Llama 4 Maverick
Long context reasoning (AA-LCR)50.0%Not measuredOnly Llama 4 Maverick measured
DTBench61.9%62.9%Qwen2.5 72B Instruct
LMCA15.9%13.4%Llama 4 Maverick
Fiction.LiveBench46.2%Not measuredOnly Llama 4 Maverick measured
Creative writing (Lech Mazur)62.0%Not measuredOnly Llama 4 Maverick measured
HellaSwagNot measured84.8%Only Qwen2.5 72B Instruct measured
WinoGrandeNot measured82.3%Only Qwen2.5 72B Instruct measured
PIQANot measured82.6%Only Qwen2.5 72B Instruct measured
LMArena Text12871269Llama 4 Maverick
LMArena Text: instruction following12671254Llama 4 Maverick
LMArena Text: creative writing12661221Llama 4 Maverick
LMArena Vision1142Not measuredOnly Llama 4 Maverick measured
LMArena Text: English13031283Llama 4 Maverick
LMArena Text: non-English12691252Llama 4 Maverick
LMArena Text: Chinese12771273Llama 4 Maverick
LMArena Text: German12911234Llama 4 Maverick
LMArena Text: Japanese12061180Llama 4 Maverick
LMArena Text: Korean12031188Llama 4 Maverick
LMArena Text: French12591280Qwen2.5 72B Instruct
LMArena Text: Spanish12931255Llama 4 Maverick
LMArena Text: Russian12861264Llama 4 Maverick
LMArena Text: Polish1275Not measuredOnly Llama 4 Maverick measured
LMArena Text: multi-turn12891272Llama 4 Maverick
LMArena Text: longer queries12801281Qwen2.5 72B Instruct
LMArena Text: writing and language12771247Llama 4 Maverick
LMArena Text: entertainment and media12501211Llama 4 Maverick
LMArena Vision: English1150Not measuredOnly Llama 4 Maverick measured
LMArena Vision: Chinese1141Not measuredOnly Llama 4 Maverick measured
LMArena Vision: reading text1143Not measuredOnly Llama 4 Maverick measured
LMArena Vision: humor1100Not measuredOnly Llama 4 Maverick measured
LMArena Vision: creative writing1099Not measuredOnly Llama 4 Maverick measured
Neither model has a result on the other 12 boards here
Image generationNot comparablePublic boards, shown as published, kept out of the index0 boards, none shared
Image generation benchmarks for Llama 4 Maverick and Qwen2.5 72B Instruct
BenchmarkLlama 4 MaverickQwen2.5 72B InstructLeader
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

Llama 4 Maverick$0.0008Fits in one request
Qwen2.5 72B Instruct$0.0006Fits in one request

Qwen2.5 72B Instruct costs less

Repository review

50K input + 3K output tokens

Llama 4 Maverick$0.021Fits in one request
Qwen2.5 72B Instruct$0.019Exceeds the context window

Qwen2.5 72B Instruct costs less

Agent loop

200K input + 10K output tokens

Llama 4 Maverick$0.080Fits in one request
Qwen2.5 72B Instruct$0.076Exceeds the context window

Qwen2.5 72B Instruct 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 Llama 4 Maverick and Qwen2.5 72B Instruct
SpecificationMeta Llama 4 MaverickQwen2.5 72B Instruct
MakerMeta LlamaQwen
WeightsOpen weightsOpen weights
ReleasedApril 5, 2025September 19, 2024
Context window524K33K
Input $/Mtok$0.35$0.36
Output $/Mtok$1.00$0.40
Coding indexNot benchmarkedNot benchmarked
Independent measurements behind it00
Cost per solved task$1.491Not computable
SourceOpenRouter listingOpenRouter listing

Questions

Is Llama 4 Maverick better than Qwen2.5 72B Instruct for coding?

Neither model has a coding score yet, so the two cannot be ranked on coding. Pick Llama 4 Maverick for a larger context window (524K against 33K tokens) and work measured by τ²-Bench Telecom or LMArena Text: legal and government, where it leads.

Which is cheaper, Llama 4 Maverick or Qwen2.5 72B Instruct?

Llama 4 Maverick costs $0.35 per million input tokens and $1.00 per million output tokens; Qwen2.5 72B Instruct costs $0.36 and $0.40. Cost per solved task is known for Llama 4 Maverick only.

Which has the bigger context window, Llama 4 Maverick or Qwen2.5 72B Instruct?

Llama 4 Maverick, with 524K tokens against 33K.

How many benchmarks is the Llama 4 Maverick vs Qwen2.5 72B Instruct comparison based on?

41 benchmarks have a result for both models as of October 22, 2026. Llama 4 Maverick's coding index rests on 0 independent benchmarks and Qwen2.5 72B Instruct's on 0. Where only one model has a result, the page says so instead of naming a winner.

Compare next