Enter opens the highlighted result. Esc closes.

Qwen3 235B A22B Thinking 2507 vs Inkling

Inkling leads Qwen3 235B A22B Thinking 2507 on the Artificial Analysis Intelligence Index, 25.0 to 12.7, at about the same output price ($4.05 vs $3.50 per million). Updated October 22, 2026; 42 benchmarks have a result for both.

Replace Qwen3 235B A22B Thinking 2507 Replace Inkling
Qwen3 235B A22B Thinking 2507Qwen
Coding
No data
Overall
52.9
Output $/M
$3.50
3
T InklingThinking Machines
Coding
31.5
Overall
46.5
Output $/M
$4.05
39
vs boards won of 42 shared

Qwen3 235B A22B Thinking 2507 or Inkling: which to pick

Only Inkling has a coding score, 31.5 on the Vibe Coding Index; Qwen3 235B A22B Thinking 2507 has none, so the two cannot be ranked on coding.

Pick Qwen3 235B A22B Thinking 2507 for output tokens 1.2x cheaper ($3.50 vs $4.05 per million) and work measured by WeirdML or SimpleQA Verified, where it leads.

Model AQwen3 235B A22B Thinking 2507QwenNot benchmarkedNo coding benchmarkNo cost per solved taskModel BInklingThinking Machines31.5/1006 independent benchmarks$0.76 per solved task
Shared results
42
Qwen3 235B A22B Thinking 2507 only
4
Inkling only
60
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
Qwen3 235B A22B Thinking 2507 not benchmarked, Inkling 31.5 from 6 benchmarks.
Widest category
Web design, 24.6 points apart on average over 6 shared benchmarks.
Widest single gap
Design Arena: UI components: Qwen3 235B A22B Thinking 2507 942 against Inkling 1198, 32.0 points apart on a 0 to 100 scale.
Output price
Qwen3 235B A22B Thinking 2507 $3.50 against Inkling $4.05 per million output tokens, 1.2x apart.
Context window
Qwen3 235B A22B Thinking 2507 takes 128K tokens, Inkling 524K.

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
CategoryQwen3 235B A22B Thinking 2507BarsInklingTies
Coding13
Intelligence01
Agentic work01
Reasoning04
Math03
Knowledge23
Web design06
General018
All shared boards339
CodingLike-for-likeSix of these boards feed the Vibe Coding Index13
Coding benchmarks for Qwen3 235B A22B Thinking 2507 and Inkling
BenchmarkQwen3 235B A22B Thinking 2507InklingLeader
LiveBench codingNot measured64.0%Only Inkling measured
LMArena WebDevNot measured1414Only Inkling measured
AA Coding Index22.152.1Inkling
LiveCodeBench (Vals AI)Not measured85.5%Only Inkling measured
SWE-bench Verified (Vals AI)Not measured77.6%Only Inkling measured
Terminal-Bench 4 (Vals AI)Not measured0.5%Only Inkling measured
Vibe Code Bench (Vals AI)Not measured19.2%Only Inkling measured
Code migration (Vals AI)Not measured11.8%Only Inkling measured
ProgramBench (Vals AI)Not measured0.0%Only Inkling measured
IOI (Vals AI)Not measured14.9%Only Inkling measured
Terminal-Bench 2.1 (Vals AI)Not measured47.6%Only Inkling measured
Vibe Code Bench 1-100 (Vals AI)Not measured7.3%Only Inkling measured
SkillsBench (Vals AI)Not measured26.1%Only Inkling measured
CyberBench (Vals AI)Not measured66.8%Only Inkling measured
SciCodeNot measured47.0%Only Inkling measured
FrontierCodeNot measured14.0%Only Inkling measured
FrontierSWENot measured4.1%Only Inkling measured
WeirdML41.0%32.3%Qwen3 235B A22B Thinking 2507
ALE-BenchNot measured946.0Only Inkling measured
LMArena Text: coding14241466Inkling
LMArena Text: software and IT14311465Inkling
LMArena WebDev: ReactNot measured1404Only Inkling measured
LMArena WebDev: HTMLNot measured1399Only Inkling measured
Neither model has a result on the other 16 boards here
IntelligenceLike-for-likePublic boards, shown as published, kept out of the index01
Intelligence benchmarks for Qwen3 235B A22B Thinking 2507 and Inkling
BenchmarkQwen3 235B A22B Thinking 2507InklingLeader
AA Intelligence Index12.725.0Inkling
Epoch Capabilities IndexNot measured148.4Only Inkling measured
Agentic workLike-for-likePublic boards, shown as published, kept out of the index01
Agentic work benchmarks for Qwen3 235B A22B Thinking 2507 and Inkling
BenchmarkQwen3 235B A22B Thinking 2507InklingLeader
AA Agentic Index1.322.5Inkling
Design Arena: agents building web appsNot measured1196Only Inkling measured
LiveBench researchNot measured59.9%Only Inkling measured
LiveBench browserNot measured16.0%Only Inkling measured
LiveBench long agenticNot measured14.1%Only Inkling measured
SREBench (Vals AI)Not measured0.0%Only Inkling measured
Finance Agent (Vals AI)Not measured46.6%Only Inkling measured
Tax Agent Bench (Vals AI)Not measured15.3%Only Inkling measured
Legal Research Bench (Vals AI)Not measured28.4%Only Inkling measured
Harvey's Legal Agent Benchmark (Vals AI)Not measured3.8%Only Inkling measured
Excel Modeling Benchmark (Vals AI)Not measured38.4%Only Inkling measured
APEX-AgentsNot measured33.8%Only Inkling measured
Vending-Bench 2-$11Not measuredOnly Qwen3 235B A22B Thinking 2507 measured
Neither model has a result on the other 28 boards here
ReasoningLike-for-likePublic boards, shown as published, kept out of the index04
Reasoning benchmarks for Qwen3 235B A22B Thinking 2507 and Inkling
BenchmarkQwen3 235B A22B Thinking 2507InklingLeader
GPQA DiamondNot measured87.1%Only Inkling measured
LiveBench reasoningNot measured78.3%Only Inkling measured
MMMU Pro (Vals AI)Not measured78.5%Only Inkling measured
GPQA Diamond (Artificial Analysis run)Not measured87.2%Only Inkling measured
Humanity's Last Exam (Artificial Analysis run)Not measured31.9%Only Inkling measured
GPQA Diamond (Epoch AI run)80.1%88.3%Inkling
ARC-AGINot measured79.5%Only Inkling measured
ARC-AGI-2Not measured36.5%Only Inkling measured
SimpleBenchNot measured50.0%Only Inkling measured
Chess puzzlesNot measured21.0%Only Inkling measured
Mystery game puzzles9.0%Not measuredOnly Qwen3 235B A22B Thinking 2507 measured
LMArena Text: hard prompts14161451Inkling
LMArena Text: expert14631466Inkling
LMArena Text: hard prompts in English14321448Inkling
Neither model has a result on the other 8 boards here
MathLike-for-likePublic boards, shown as published, kept out of the index03
Math benchmarks for Qwen3 235B A22B Thinking 2507 and Inkling
BenchmarkQwen3 235B A22B Thinking 2507InklingLeader
OTIS Mock AIME86.7%88.9%Inkling
FrontierMath, tiers 1 to 3Not measured33.3%Only Inkling measured
FrontierMath, tier 4Not measured4.9%Only Inkling measured
ProofBenchNot measured0.0%Only Inkling measured
LMArena Text: math14121479Inkling
LMArena Text: mathematical work14351460Inkling
Neither model has a result on the other 7 boards here
KnowledgeLike-for-likePublic boards, shown as published, kept out of the index23
Knowledge benchmarks for Qwen3 235B A22B Thinking 2507 and Inkling
BenchmarkQwen3 235B A22B Thinking 2507InklingLeader
MMLU ProNot measured86.3%Only Inkling measured
LegalBench (Vals AI)Not measured83.0%Only Inkling measured
CorpFin (Vals AI)Not measured68.6%Only Inkling measured
TaxEval (Vals AI)Not measured75.3%Only Inkling measured
MortgageTax (Vals AI)Not measured64.0%Only Inkling measured
MedCode (Vals AI)Not measured41.2%Only Inkling measured
SimpleQA Verified40.4%40.3%Qwen3 235B A22B Thinking 2507
LMArena Text: science14331455Inkling
LMArena Text: medicine and healthcare14561447Qwen3 235B A22B Thinking 2507
LMArena Text: legal and government14151448Inkling
LMArena Text: business and finance14131438Inkling
Neither model has a result on the other 7 boards here
Web designLike-for-likePublic boards, shown as published, kept out of the index06
Web design benchmarks for Qwen3 235B A22B Thinking 2507 and Inkling
BenchmarkQwen3 235B A22B Thinking 2507InklingLeader
Design Arena: websites10521224Inkling
Design Arena: UI components9421198Inkling
Design Arena: game dev9561165Inkling
Design Arena: data viz9501156Inkling
Design Arena: 3D9971160Inkling
Design Arena: all code categories10321205Inkling
Design Arena: SVGNot measured1108Only Inkling measured
Design Arena: ASCII artNot measured1089Only Inkling measured
LiveBench designNot measured58.3%Only Inkling measured
GeneralLike-for-likePublic boards, shown as published, kept out of the index018
General benchmarks for Qwen3 235B A22B Thinking 2507 and Inkling
BenchmarkQwen3 235B A22B Thinking 2507InklingLeader
Vals IndexNot measured29.0%Only Inkling measured
LiveBench data analysisNot measured72.8%Only Inkling measured
Public Benefits Bench (Vals AI)Not measured58.5%Only Inkling measured
SAGE (Vals AI)Not measured36.5%Only Inkling measured
MedScribe (Vals AI)Not measured85.4%Only Inkling measured
Vals Multimodal IndexNot measured49.4%Only Inkling measured
Long context reasoning (AA-LCR)Not measured77.3%Only Inkling measured
DTBench80.3%87.5%Inkling
LMCA29.3%37.6%Inkling
Fiction.LiveBench75.0%Not measuredOnly Qwen3 235B A22B Thinking 2507 measured
Creative writing (Lech Mazur)82.4%Not measuredOnly Qwen3 235B A22B Thinking 2507 measured
LMArena Text14151442Inkling
LMArena Text: instruction following13861425Inkling
LMArena Text: creative writing13841390Inkling
LMArena Text: English14291442Inkling
LMArena Text: non-English13971434Inkling
LMArena Text: Chinese14811493Inkling
LMArena Text: German14081445Inkling
LMArena Text: Japanese13861429Inkling
LMArena Text: Korean13751403Inkling
LMArena Text: FrenchNot measured1455Only Inkling measured
LMArena Text: Spanish13891449Inkling
LMArena Text: Russian14011430Inkling
LMArena Text: Polish13901450Inkling
LMArena Text: multi-turn14061437Inkling
LMArena Text: longer queries14001435Inkling
LMArena Text: writing and language13881404Inkling
LMArena Text: entertainment and media13701394Inkling
Neither model has a result on the other 18 boards here
Image generationNot comparablePublic boards, shown as published, kept out of the index0 boards, none shared
Image generation benchmarks for Qwen3 235B A22B Thinking 2507 and Inkling
BenchmarkQwen3 235B A22B Thinking 2507InklingLeader
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

Qwen3 235B A22B Thinking 2507$0.0022Fits in one request
Inkling$0.0030Fits in one request

Qwen3 235B A22B Thinking 2507 costs less

Repository review

50K input + 3K output tokens

Qwen3 235B A22B Thinking 2507$0.033Fits in one request
Inkling$0.061Fits in one request

Qwen3 235B A22B Thinking 2507 costs less

Agent loop

200K input + 10K output tokens

Qwen3 235B A22B Thinking 2507$0.125Exceeds the context window
Inkling$0.235Fits in one request

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 Qwen3 235B A22B Thinking 2507 and Inkling
SpecificationQwen3 235B A22B Thinking 2507Thinking Machines Inkling
MakerQwenThinking Machines
WeightsOpen weightsOpen weights
ReleasedJuly 25, 2025July 17, 2026
Context window128K524K
Input $/Mtok$0.45$0.975
Output $/Mtok$3.50$4.05
Coding indexNot benchmarked31.5
Independent measurements behind it06
Cost per solved taskNot computable$0.76
SourceOpenRouter listingOpenRouter listing

Questions

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

Only Inkling has a coding score, 31.5 on the Vibe Coding Index; Qwen3 235B A22B Thinking 2507 has none, so the two cannot be ranked on coding. Pick Qwen3 235B A22B Thinking 2507 for output tokens 1.2x cheaper ($3.50 vs $4.05 per million) and work measured by WeirdML or SimpleQA Verified, where it leads.

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

Qwen3 235B A22B Thinking 2507 costs $0.45 per million input tokens and $3.50 per million output tokens; Inkling costs $0.975 and $4.05. Cost per solved task is known for Inkling only.

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

Inkling, with 524K tokens against 128K.

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

42 benchmarks have a result for both models as of October 22, 2026. Qwen3 235B A22B Thinking 2507's coding index rests on 0 independent benchmarks and Inkling's on 6. Where only one model has a result, the page says so instead of naming a winner.

Compare next