Enter opens the highlighted result. Esc closes.

GPT-4o (2024-08-06) vs Qwen3 235B A22B Thinking 2507

Qwen3 235B A22B Thinking 2507 leads GPT-4o (2024-08-06) on the Artificial Analysis Intelligence Index, 12.7 to 7.7, and Qwen3 235B A22B Thinking 2507 costs 2.9x less per output token ($3.50 vs $10.00 per million). Updated October 23, 2026; 30 benchmarks have a result for both.

Replace GPT-4o (2024-08-06) Replace Qwen3 235B A22B Thinking 2507
GPT-4o (2024-08-06)OpenAI
Coding
No data
Overall
26.3
Output $/M
$10.00
0
Qwen3 235B A22B Thinking 2507Qwen
Coding
No data
Overall
52.9
Output $/M
$3.50
30
vs boards won of 30 shared

GPT-4o (2024-08-06) 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 GPT-4o (2024-08-06) for work measured by LMArena Text: coding or LMArena Text: software and IT, where it leads.

Model AGPT-4o (2024-08-06)OpenAINot benchmarkedNo coding benchmarkNo cost per solved taskModel BQwen3 235B A22B Thinking 2507QwenNot benchmarkedNo coding benchmarkNo cost per solved task
Shared results
30
GPT-4o (2024-08-06) only
28
Qwen3 235B A22B Thinking 2507 only
16
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
GPT-4o (2024-08-06) not benchmarked, Qwen3 235B A22B Thinking 2507 not benchmarked.
Widest category
Math, 42.2 points apart on average over 3 shared benchmarks.
Widest single gap
OTIS Mock AIME: GPT-4o (2024-08-06) 6.4% against Qwen3 235B A22B Thinking 2507 86.7%, 80.3 points apart on a 0 to 100 scale.
Output price
GPT-4o (2024-08-06) $10.00 against Qwen3 235B A22B Thinking 2507 $3.50 per million output tokens, 2.9x apart.
Context window
Both take 128K tokens.

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
CategoryGPT-4o (2024-08-06)BarsQwen3 235B A22B Thinking 2507Ties
Coding02
Intelligence01
Reasoning04
Math03
Knowledge05
General015
All shared boards030
CodingLike-for-likeSix of these boards feed the Vibe Coding Index02
Coding benchmarks for GPT-4o (2024-08-06) and Qwen3 235B A22B Thinking 2507
BenchmarkGPT-4o (2024-08-06)Qwen3 235B A22B Thinking 2507Leader
Aider polyglot23.1%Not measuredOnly GPT-4o (2024-08-06) measured
SWE-bench Lite21.7%Not measuredOnly GPT-4o (2024-08-06) measured
AA Coding IndexNot measured22.1Only Qwen3 235B A22B Thinking 2507 measured
LiveCodeBench (Artificial Analysis run)31.7%Not measuredOnly GPT-4o (2024-08-06) measured
Terminal-Bench Hard (Artificial Analysis run)8.3%Not measuredOnly GPT-4o (2024-08-06) measured
WeirdMLNot measured41.0%Only Qwen3 235B A22B Thinking 2507 measured
LMArena Text: coding12831424Qwen3 235B A22B Thinking 2507
LMArena Text: software and IT12781431Qwen3 235B A22B Thinking 2507
Neither model has a result on the other 35 boards here
IntelligenceLike-for-likePublic boards, shown as published, kept out of the index01
Intelligence benchmarks for GPT-4o (2024-08-06) and Qwen3 235B A22B Thinking 2507
BenchmarkGPT-4o (2024-08-06)Qwen3 235B A22B Thinking 2507Leader
AA Intelligence Index7.712.7Qwen3 235B A22B Thinking 2507
Neither model has a result on the other 1 board here
Agentic workDirectional onlyPublic boards, shown as published, kept out of the index3 boards, none shared
Agentic work benchmarks for GPT-4o (2024-08-06) and Qwen3 235B A22B Thinking 2507
BenchmarkGPT-4o (2024-08-06)Qwen3 235B A22B Thinking 2507Leader
AA Agentic IndexNot measured1.3Only Qwen3 235B A22B Thinking 2507 measured
τ²-Bench Telecom28.9%Not measuredOnly GPT-4o (2024-08-06) measured
Vending-Bench 2Not measured-$11Only Qwen3 235B A22B Thinking 2507 measured
Neither model has a result on the other 39 boards here
ReasoningLike-for-likePublic boards, shown as published, kept out of the index04
Reasoning benchmarks for GPT-4o (2024-08-06) and Qwen3 235B A22B Thinking 2507
BenchmarkGPT-4o (2024-08-06)Qwen3 235B A22B Thinking 2507Leader
MMMU Pro (Vals AI)64.0%Not measuredOnly GPT-4o (2024-08-06) measured
GPQA Diamond (Artificial Analysis run)52.1%Not measuredOnly GPT-4o (2024-08-06) measured
Humanity's Last Exam (Artificial Analysis run)2.3%Not measuredOnly GPT-4o (2024-08-06) measured
GPQA Diamond (Epoch AI run)49.2%80.1%Qwen3 235B A22B Thinking 2507
SimpleBench17.8%Not measuredOnly GPT-4o (2024-08-06) measured
Chess puzzles6.0%Not measuredOnly GPT-4o (2024-08-06) measured
Mystery game puzzlesNot measured9.0%Only Qwen3 235B A22B Thinking 2507 measured
LMArena Text: hard prompts12641416Qwen3 235B A22B Thinking 2507
LMArena Text: expert12411463Qwen3 235B A22B Thinking 2507
LMArena Text: hard prompts in English12771432Qwen3 235B A22B Thinking 2507
Neither model has a result on the other 13 boards here
MathLike-for-likePublic boards, shown as published, kept out of the index03
Math benchmarks for GPT-4o (2024-08-06) and Qwen3 235B A22B Thinking 2507
BenchmarkGPT-4o (2024-08-06)Qwen3 235B A22B Thinking 2507Leader
AIME14.0%Not measuredOnly GPT-4o (2024-08-06) measured
MATH 500 (Vals AI)75.2%Not measuredOnly GPT-4o (2024-08-06) measured
MGSM (Vals AI)90.7%Not measuredOnly GPT-4o (2024-08-06) measured
OTIS Mock AIME6.4%86.7%Qwen3 235B A22B Thinking 2507
FrontierMath, tiers 1 to 30.4%Not measuredOnly GPT-4o (2024-08-06) measured
MATH level 553.3%Not measuredOnly GPT-4o (2024-08-06) measured
LMArena Text: math12851412Qwen3 235B A22B Thinking 2507
LMArena Text: mathematical work12831435Qwen3 235B A22B Thinking 2507
Neither model has a result on the other 6 boards here
KnowledgeLike-for-likePublic boards, shown as published, kept out of the index05
Knowledge benchmarks for GPT-4o (2024-08-06) and Qwen3 235B A22B Thinking 2507
BenchmarkGPT-4o (2024-08-06)Qwen3 235B A22B Thinking 2507Leader
MMLU Pro74.1%Not measuredOnly GPT-4o (2024-08-06) measured
LegalBench (Vals AI)80.1%Not measuredOnly GPT-4o (2024-08-06) measured
CorpFin (Vals AI)39.4%Not measuredOnly GPT-4o (2024-08-06) measured
TaxEval (Vals AI)71.1%Not measuredOnly GPT-4o (2024-08-06) measured
MortgageTax (Vals AI)61.0%Not measuredOnly GPT-4o (2024-08-06) measured
MedQA (Vals AI)88.2%Not measuredOnly GPT-4o (2024-08-06) measured
SimpleQA Verified26.0%40.4%Qwen3 235B A22B Thinking 2507
MMLU84.3%Not measuredOnly GPT-4o (2024-08-06) measured
LMArena Text: science12741433Qwen3 235B A22B Thinking 2507
LMArena Text: medicine and healthcare12521456Qwen3 235B A22B Thinking 2507
LMArena Text: legal and government12751415Qwen3 235B A22B Thinking 2507
LMArena Text: business and finance12321412Qwen3 235B A22B Thinking 2507
Neither model has a result on the other 8 boards here
Web designDirectional onlyPublic boards, shown as published, kept out of the index6 boards, none shared
Web design benchmarks for GPT-4o (2024-08-06) and Qwen3 235B A22B Thinking 2507
BenchmarkGPT-4o (2024-08-06)Qwen3 235B A22B Thinking 2507Leader
Design Arena: websitesNot measured1052Only Qwen3 235B A22B Thinking 2507 measured
Design Arena: UI componentsNot measured942Only Qwen3 235B A22B Thinking 2507 measured
Design Arena: game devNot measured956Only Qwen3 235B A22B Thinking 2507 measured
Design Arena: data vizNot measured950Only Qwen3 235B A22B Thinking 2507 measured
Design Arena: 3DNot measured998Only Qwen3 235B A22B Thinking 2507 measured
Design Arena: all code categoriesNot measured1032Only Qwen3 235B A22B Thinking 2507 measured
Neither model has a result on the other 3 boards here
GeneralLike-for-likePublic boards, shown as published, kept out of the index015
General benchmarks for GPT-4o (2024-08-06) and Qwen3 235B A22B Thinking 2507
BenchmarkGPT-4o (2024-08-06)Qwen3 235B A22B Thinking 2507Leader
IFBench36.0%Not measuredOnly GPT-4o (2024-08-06) measured
Long context reasoning (AA-LCR)41.0%Not measuredOnly GPT-4o (2024-08-06) measured
DTBenchNot measured80.3%Only Qwen3 235B A22B Thinking 2507 measured
LMCANot measured29.3%Only Qwen3 235B A22B Thinking 2507 measured
Fiction.LiveBenchNot measured75.0%Only Qwen3 235B A22B Thinking 2507 measured
Creative writing (Lech Mazur)Not measured82.4%Only Qwen3 235B A22B Thinking 2507 measured
LMArena Text12831415Qwen3 235B A22B Thinking 2507
LMArena Text: instruction following12671386Qwen3 235B A22B Thinking 2507
LMArena Text: creative writing12741384Qwen3 235B A22B Thinking 2507
LMArena Vision1065Not measuredOnly GPT-4o (2024-08-06) measured
LMArena Text: English12971429Qwen3 235B A22B Thinking 2507
LMArena Text: non-English12621397Qwen3 235B A22B Thinking 2507
LMArena Text: Chinese12541481Qwen3 235B A22B Thinking 2507
LMArena Text: German12571408Qwen3 235B A22B Thinking 2507
LMArena Text: Japanese12341386Qwen3 235B A22B Thinking 2507
LMArena Text: Korean12181375Qwen3 235B A22B Thinking 2507
LMArena Text: French1263Not measuredOnly GPT-4o (2024-08-06) measured
LMArena Text: Spanish12691389Qwen3 235B A22B Thinking 2507
LMArena Text: Russian12721400Qwen3 235B A22B Thinking 2507
LMArena Text: PolishNot measured1390Only Qwen3 235B A22B Thinking 2507 measured
LMArena Text: multi-turn12791406Qwen3 235B A22B Thinking 2507
LMArena Text: longer queries12831399Qwen3 235B A22B Thinking 2507
LMArena Text: writing and language12901388Qwen3 235B A22B Thinking 2507
LMArena Text: entertainment and media12671370Qwen3 235B A22B Thinking 2507
LMArena Vision: English1072Not measuredOnly GPT-4o (2024-08-06) measured
LMArena Vision: Chinese1009Not measuredOnly GPT-4o (2024-08-06) measured
Neither model has a result on the other 22 boards here
Image generationNot comparablePublic boards, shown as published, kept out of the index0 boards, none shared
Image generation benchmarks for GPT-4o (2024-08-06) and Qwen3 235B A22B Thinking 2507
BenchmarkGPT-4o (2024-08-06)Qwen3 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

GPT-4o (2024-08-06)$0.0075Fits 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

GPT-4o (2024-08-06)$0.155Fits 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

GPT-4o (2024-08-06)$0.600Exceeds 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 GPT-4o (2024-08-06) and Qwen3 235B A22B Thinking 2507
SpecificationOpenAI GPT-4o (2024-08-06)Qwen3 235B A22B Thinking 2507
MakerOpenAIQwen
WeightsNo public weights listedOpen weights
ReleasedAugust 6, 2024July 25, 2025
Context window128K128K
Input $/Mtok$2.50$0.45
Output $/Mtok$10.00$3.50
Coding indexNot benchmarkedNot benchmarked
Independent measurements behind it00
Cost per solved taskNot computableNot computable
SourceOpenRouter listingOpenRouter listing

Questions

Is GPT-4o (2024-08-06) 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 GPT-4o (2024-08-06) for work measured by LMArena Text: coding or LMArena Text: software and IT, where it leads.

Which is cheaper, GPT-4o (2024-08-06) or Qwen3 235B A22B Thinking 2507?

GPT-4o (2024-08-06) costs $2.50 per million input tokens and $10.00 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, GPT-4o (2024-08-06) or Qwen3 235B A22B Thinking 2507?

Neither: both take 128K tokens.

How many benchmarks is the GPT-4o (2024-08-06) vs Qwen3 235B A22B Thinking 2507 comparison based on?

30 benchmarks have a result for both models as of October 23, 2026. GPT-4o (2024-08-06)'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