Enter opens the highlighted result. Esc closes.

o4 Mini vs Qwen3 235B A22B Thinking 2507

o4 Mini leads Qwen3 235B A22B Thinking 2507 on the Artificial Analysis Intelligence Index, 16.7 to 12.7, and Qwen3 235B A22B Thinking 2507 costs 1.3x less per output token ($3.50 vs $4.40 per million). Updated October 22, 2026; 43 benchmarks have a result for both.

Replace o4 Mini Replace Qwen3 235B A22B Thinking 2507
o4 MiniOpenAI
Coding
No data
Overall
43.0
Output $/M
$4.40
6
Qwen3 235B A22B Thinking 2507Qwen
Coding
No data
Overall
52.9
Output $/M
$3.50
37
vs boards won of 43 shared

o4 Mini or Qwen3 235B A22B Thinking 2507: which to pick

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

Pick o4 Mini for a larger context window (200K against 128K tokens) and work measured by LMArena Text: coding or LMArena Text: software and IT, where it leads.

Model Ao4 MiniOpenAINot benchmarkedNo coding benchmarkNo cost per solved taskModel BQwen3 235B A22B Thinking 2507QwenNot benchmarkedNo coding benchmarkNo cost per solved task
Shared results
43
o4 Mini only
55
Qwen3 235B A22B Thinking 2507 only
3
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
o4 Mini not benchmarked, Qwen3 235B A22B Thinking 2507 not benchmarked.
Widest category
Knowledge, 15.1 points apart on average over 5 shared benchmarks.
Widest single gap
LMArena Text: Chinese: o4 Mini 1354 against Qwen3 235B A22B Thinking 2507 1481, 21.1 points apart on a 0 to 100 scale.
Output price
o4 Mini $4.40 against Qwen3 235B A22B Thinking 2507 $3.50 per million output tokens, 1.3x apart.
Context window
o4 Mini takes 200K tokens, Qwen3 235B A22B Thinking 2507 128K.

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
Categoryo4 MiniBarsQwen3 235B A22B Thinking 2507Ties
Coding12
Intelligence10
Reasoning05
Math03
Knowledge05
Web design33
General119
All shared boards637
CodingLike-for-likeSix of these boards feed the Vibe Coding Index12
Coding benchmarks for o4 Mini and Qwen3 235B A22B Thinking 2507
Benchmarko4 MiniQwen3 235B A22B Thinking 2507Leader
Aider polyglot72.0%Not measuredOnly o4 Mini measured
SWE-bench Verified64.6%Not measuredOnly o4 Mini measured
SWE-bench Multimodal33.9%Not measuredOnly o4 Mini measured
AA Coding IndexNot measured22.1Only Qwen3 235B A22B Thinking 2507 measured
LiveCodeBench (Vals AI)82.2%Not measuredOnly o4 Mini measured
LiveCodeBench (Artificial Analysis run)85.9%Not measuredOnly o4 Mini measured
Terminal-Bench Hard (Artificial Analysis run)15.2%Not measuredOnly o4 Mini measured
GSO-Bench3.6%Not measuredOnly o4 Mini measured
WeirdML52.6%41.0%o4 Mini
LMArena Text: coding13681424Qwen3 235B A22B Thinking 2507
LMArena Text: software and IT13671431Qwen3 235B A22B Thinking 2507
SciCode10.8%Not measuredOnly o4 Mini measured
Neither model has a result on the other 27 boards here
IntelligenceLike-for-likePublic boards, shown as published, kept out of the index10
Intelligence benchmarks for o4 Mini and Qwen3 235B A22B Thinking 2507
Benchmarko4 MiniQwen3 235B A22B Thinking 2507Leader
AA Intelligence Index16.712.7o4 Mini
Epoch Capabilities Index145.5Not measuredOnly o4 Mini measured
Agentic workDirectional onlyPublic boards, shown as published, kept out of the index5 boards, none shared
Agentic work benchmarks for o4 Mini and Qwen3 235B A22B Thinking 2507
Benchmarko4 MiniQwen3 235B A22B Thinking 2507Leader
AA Agentic IndexNot measured1.3Only Qwen3 235B A22B Thinking 2507 measured
τ²-Bench Telecom55.6%Not measuredOnly o4 Mini measured
METR time horizons63.9%Not measuredOnly o4 Mini measured
Vending-Bench 2Not measured-$11Only Qwen3 235B A22B Thinking 2507 measured
Berkeley Function Calling Leaderboard (BFCL)53.2%Not measuredOnly o4 Mini measured
Neither model has a result on the other 36 boards here
ReasoningLike-for-likePublic boards, shown as published, kept out of the index05
Reasoning benchmarks for o4 Mini and Qwen3 235B A22B Thinking 2507
Benchmarko4 MiniQwen3 235B A22B Thinking 2507Leader
GPQA Diamond74.5%Not measuredOnly o4 Mini measured
MMMU Pro (Vals AI)79.7%Not measuredOnly o4 Mini measured
GPQA Diamond (Artificial Analysis run)78.4%Not measuredOnly o4 Mini measured
Humanity's Last Exam (Artificial Analysis run)16.5%Not measuredOnly o4 Mini measured
GPQA Diamond (Epoch AI run)79.6%80.1%Qwen3 235B A22B Thinking 2507
ARC-AGI58.7%Not measuredOnly o4 Mini measured
ARC-AGI-26.1%Not measuredOnly o4 Mini measured
SimpleBench38.7%Not measuredOnly o4 Mini measured
Chess puzzles22.0%Not measuredOnly o4 Mini measured
Mystery game puzzles5.0%9.0%Qwen3 235B A22B Thinking 2507
VPCT57.5%Not measuredOnly o4 Mini measured
LMArena Text: hard prompts13511416Qwen3 235B A22B Thinking 2507
LMArena Text: expert13431463Qwen3 235B A22B Thinking 2507
LMArena Text: hard prompts in English13661432Qwen3 235B A22B Thinking 2507
LMArena Vision: diagrams1197Not measuredOnly o4 Mini measured
LMArena Vision: homework1234Not measuredOnly o4 Mini measured
Kagi LLM Benchmark67.6%Not measuredOnly o4 Mini measured
Neither model has a result on the other 5 boards here
MathLike-for-likePublic boards, shown as published, kept out of the index03
Math benchmarks for o4 Mini and Qwen3 235B A22B Thinking 2507
Benchmarko4 MiniQwen3 235B A22B Thinking 2507Leader
AIME83.7%Not measuredOnly o4 Mini measured
MATH 500 (Vals AI)94.2%Not measuredOnly o4 Mini measured
MGSM (Vals AI)93.4%Not measuredOnly o4 Mini measured
AA Math Index90.7Not measuredOnly o4 Mini measured
AIME 2025 (Artificial Analysis run)90.7%Not measuredOnly o4 Mini measured
OTIS Mock AIME81.7%86.7%Qwen3 235B A22B Thinking 2507
FrontierMath, tiers 1 to 336.1%Not measuredOnly o4 Mini measured
FrontierMath, tier 44.9%Not measuredOnly o4 Mini measured
MATH level 597.8%Not measuredOnly o4 Mini measured
LMArena Text: math13891412Qwen3 235B A22B Thinking 2507
LMArena Text: mathematical work13791435Qwen3 235B A22B Thinking 2507
Neither model has a result on the other 2 boards here
KnowledgeLike-for-likePublic boards, shown as published, kept out of the index05
Knowledge benchmarks for o4 Mini and Qwen3 235B A22B Thinking 2507
Benchmarko4 MiniQwen3 235B A22B Thinking 2507Leader
MMLU Pro80.6%Not measuredOnly o4 Mini measured
LegalBench (Vals AI)79.2%Not measuredOnly o4 Mini measured
CorpFin (Vals AI)59.0%Not measuredOnly o4 Mini measured
TaxEval (Vals AI)74.8%Not measuredOnly o4 Mini measured
MortgageTax (Vals AI)64.8%Not measuredOnly o4 Mini measured
MedQA (Vals AI)96.0%Not measuredOnly o4 Mini measured
MedCode (Vals AI)33.8%Not measuredOnly o4 Mini measured
MMLU Pro (Artificial Analysis run)83.2%Not measuredOnly o4 Mini measured
SimpleQA Verified19.6%40.4%Qwen3 235B A22B Thinking 2507
GeoBench64.0%Not measuredOnly o4 Mini measured
LMArena Text: science13541433Qwen3 235B A22B Thinking 2507
LMArena Text: medicine and healthcare13611456Qwen3 235B A22B Thinking 2507
LMArena Text: legal and government13501415Qwen3 235B A22B Thinking 2507
LMArena Text: business and finance13251413Qwen3 235B A22B Thinking 2507
FACTS Benchmark Suite28.1%Not measuredOnly o4 Mini measured
Neither model has a result on the other 3 boards here
Web designLike-for-likePublic boards, shown as published, kept out of the index33
Web design benchmarks for o4 Mini and Qwen3 235B A22B Thinking 2507
Benchmarko4 MiniQwen3 235B A22B Thinking 2507Leader
Design Arena: websites9851052Qwen3 235B A22B Thinking 2507
Design Arena: UI components983942o4 Mini
Design Arena: game dev1003956o4 Mini
Design Arena: data viz991950o4 Mini
Design Arena: 3D854997Qwen3 235B A22B Thinking 2507
Design Arena: all code categories9761032Qwen3 235B A22B Thinking 2507
Neither model has a result on the other 3 boards here
GeneralLike-for-likePublic boards, shown as published, kept out of the index119
General benchmarks for o4 Mini and Qwen3 235B A22B Thinking 2507
Benchmarko4 MiniQwen3 235B A22B Thinking 2507Leader
SAGE (Vals AI)41.1%Not measuredOnly o4 Mini measured
MedScribe (Vals AI)69.1%Not measuredOnly o4 Mini measured
IFBench68.7%Not measuredOnly o4 Mini measured
Long context reasoning (AA-LCR)61.0%Not measuredOnly o4 Mini measured
DTBench77.6%80.3%Qwen3 235B A22B Thinking 2507
LMCA26.5%29.3%Qwen3 235B A22B Thinking 2507
Fiction.LiveBench77.8%75.0%o4 Mini
Creative writing (Lech Mazur)75.0%82.4%Qwen3 235B A22B Thinking 2507
LMArena Text13531415Qwen3 235B A22B Thinking 2507
LMArena Text: instruction following13201386Qwen3 235B A22B Thinking 2507
LMArena Text: creative writing12941384Qwen3 235B A22B Thinking 2507
LMArena Vision1194Not measuredOnly o4 Mini measured
LMArena Text: English13671429Qwen3 235B A22B Thinking 2507
LMArena Text: non-English13371397Qwen3 235B A22B Thinking 2507
LMArena Text: Chinese13541481Qwen3 235B A22B Thinking 2507
LMArena Text: German13361408Qwen3 235B A22B Thinking 2507
LMArena Text: Japanese13081386Qwen3 235B A22B Thinking 2507
LMArena Text: Korean13121375Qwen3 235B A22B Thinking 2507
LMArena Text: French1364Not measuredOnly o4 Mini measured
LMArena Text: Spanish13461389Qwen3 235B A22B Thinking 2507
LMArena Text: Russian13341401Qwen3 235B A22B Thinking 2507
LMArena Text: Polish13581390Qwen3 235B A22B Thinking 2507
LMArena Text: multi-turn13501406Qwen3 235B A22B Thinking 2507
LMArena Text: longer queries13151400Qwen3 235B A22B Thinking 2507
LMArena Text: writing and language13161388Qwen3 235B A22B Thinking 2507
LMArena Text: entertainment and media13081370Qwen3 235B A22B Thinking 2507
LMArena Vision: English1200Not measuredOnly o4 Mini measured
LMArena Vision: Chinese1218Not measuredOnly o4 Mini measured
LMArena Vision: reading text1197Not measuredOnly o4 Mini measured
LMArena Vision: humor1199Not measuredOnly o4 Mini measured
LMArena Vision: captioning1204Not measuredOnly o4 Mini measured
LMArena Vision: recognizing things1218Not measuredOnly o4 Mini measured
LMArena Vision: creative writing1166Not measuredOnly o4 Mini 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 o4 Mini and Qwen3 235B A22B Thinking 2507
Benchmarko4 MiniQwen3 235B A22B Thinking 2507Leader
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

o4 Mini$0.0033Fits in one request
Qwen3 235B A22B Thinking 2507$0.0022Fits in one request

Qwen3 235B A22B Thinking 2507 costs less

Repository review

50K input + 3K output tokens

o4 Mini$0.068Fits in one request
Qwen3 235B A22B Thinking 2507$0.033Fits in one request

Qwen3 235B A22B Thinking 2507 costs less

Agent loop

200K input + 10K output tokens

o4 Mini$0.264Exceeds the context window
Qwen3 235B A22B Thinking 2507$0.125Exceeds the context window

Qwen3 235B A22B Thinking 2507 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 o4 Mini and Qwen3 235B A22B Thinking 2507
SpecificationOpenAI o4 MiniQwen3 235B A22B Thinking 2507
MakerOpenAIQwen
WeightsNo public weights listedOpen weights
ReleasedApril 16, 2025July 25, 2025
Context window200K128K
Input $/Mtok$1.10$0.45
Output $/Mtok$4.40$3.50
Coding indexNot benchmarkedNot benchmarked
Independent measurements behind it00
Cost per solved taskNot computableNot computable
SourceOpenRouter listingOpenRouter listing

Questions

Is o4 Mini better than Qwen3 235B A22B Thinking 2507 for coding?

Neither model has a coding score yet, so the two cannot be ranked on coding. Pick o4 Mini for a larger context window (200K against 128K tokens) and work measured by LMArena Text: coding or LMArena Text: software and IT, where it leads.

Which is cheaper, o4 Mini or Qwen3 235B A22B Thinking 2507?

o4 Mini costs $1.10 per million input tokens and $4.40 per million output tokens; Qwen3 235B A22B Thinking 2507 costs $0.45 and $3.50. Cost per solved task is known for neither of them.

Which has the bigger context window, o4 Mini or Qwen3 235B A22B Thinking 2507?

o4 Mini, with 200K tokens against 128K.

How many benchmarks is the o4 Mini vs Qwen3 235B A22B Thinking 2507 comparison based on?

43 benchmarks have a result for both models as of October 22, 2026. o4 Mini's coding index rests on 0 independent benchmarks and Qwen3 235B A22B Thinking 2507's on 0. Where only one model has a result, the page says so instead of naming a winner.

Compare next