Enter opens the highlighted result. Esc closes.

GLM 4.6 vs Qwen3 235B A22B Thinking 2507

GLM 4.6 leads Qwen3 235B A22B Thinking 2507 on the Artificial Analysis Intelligence Index, 18.5 to 12.7, and GLM 4.6 costs 1.6x less per output token ($2.20 vs $3.50 per million). Updated October 23, 2026; 35 benchmarks have a result for both.

Replace GLM 4.6 Replace Qwen3 235B A22B Thinking 2507
GLM 4.6Z.ai
Coding
8.8
Overall
33.5
Output $/M
$2.20
33
Qwen3 235B A22B Thinking 2507Qwen
Coding
No data
Overall
52.9
Output $/M
$3.50
2
vs boards won of 35 shared

GLM 4.6 or Qwen3 235B A22B Thinking 2507: which to pick

Only GLM 4.6 has a coding score, 8.8 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 work measured by LMArena Text: expert or LMArena Text: medicine and healthcare, where it leads.

Model AGLM 4.6Z.ai8.8/1002 independent benchmarksNo cost per solved taskModel BQwen3 235B A22B Thinking 2507QwenNot benchmarkedNo coding benchmarkNo cost per solved task
Shared results
35
GLM 4.6 only
35
Qwen3 235B A22B Thinking 2507 only
11
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
GLM 4.6 8.8 from 2 benchmarks, Qwen3 235B A22B Thinking 2507 not benchmarked.
Widest category
Web design, 20.7 points apart on average over 6 shared benchmarks.
Widest single gap
Design Arena: data viz: GLM 4.6 1161 against Qwen3 235B A22B Thinking 2507 950, 26.4 points apart on a 0 to 100 scale.
Output price
GLM 4.6 $2.20 against Qwen3 235B A22B Thinking 2507 $3.50 per million output tokens, 1.6x apart.
Context window
GLM 4.6 takes 205K 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
CategoryGLM 4.6BarsQwen3 235B A22B Thinking 2507Ties
Coding30
Intelligence10
Reasoning21
Math20
Knowledge31
Web design60
General160
All shared boards332
CodingLike-for-likeSix of these boards feed the Vibe Coding Index30
Coding benchmarks for GLM 4.6 and Qwen3 235B A22B Thinking 2507
BenchmarkGLM 4.6Qwen3 235B A22B Thinking 2507Leader
SWE-bench Verified68.2%Not measuredOnly GLM 4.6 measured
LMArena WebDev1341Not measuredOnly GLM 4.6 measured
AA Coding Index45.822.1GLM 4.6
LiveCodeBench (Vals AI)81.0%Not measuredOnly GLM 4.6 measured
Vibe Code Bench (Vals AI)3.1%Not measuredOnly GLM 4.6 measured
LiveCodeBench (Artificial Analysis run)69.5%Not measuredOnly GLM 4.6 measured
Terminal-Bench Hard (Artificial Analysis run)28.8%Not measuredOnly GLM 4.6 measured
Terminal-Bench24.5%Not measuredOnly GLM 4.6 measured
WeirdMLNot measured41.0%Only Qwen3 235B A22B Thinking 2507 measured
ALE-Bench340.8Not measuredOnly GLM 4.6 measured
LMArena Text: coding14491424GLM 4.6
LMArena Text: software and IT14551431GLM 4.6
LMArena WebDev: HTML1350Not measuredOnly GLM 4.6 measured
Neither model has a result on the other 30 boards here
IntelligenceLike-for-likePublic boards, shown as published, kept out of the index10
Intelligence benchmarks for GLM 4.6 and Qwen3 235B A22B Thinking 2507
BenchmarkGLM 4.6Qwen3 235B A22B Thinking 2507Leader
AA Intelligence Index18.512.7GLM 4.6
Neither model has a result on the other 1 board here
Agentic workDirectional onlyPublic boards, shown as published, kept out of the index9 boards, none shared
Agentic work benchmarks for GLM 4.6 and Qwen3 235B A22B Thinking 2507
BenchmarkGLM 4.6Qwen3 235B A22B Thinking 2507Leader
AA Agentic IndexNot measured1.3Only Qwen3 235B A22B Thinking 2507 measured
Design Arena: agents building full-stack apps1029Not measuredOnly GLM 4.6 measured
Design Arena: agents building mobile apps1096Not measuredOnly GLM 4.6 measured
Design Arena: agents building Android apps1086Not measuredOnly GLM 4.6 measured
Design Arena: agents building Godot games1180Not measuredOnly GLM 4.6 measured
τ²-Bench Telecom76.9%Not measuredOnly GLM 4.6 measured
Vending-Bench 2Not measured-$11Only Qwen3 235B A22B Thinking 2507 measured
Berkeley Function Calling Leaderboard (BFCL)72.4%Not measuredOnly GLM 4.6 measured
BrowseComp-Plus48.2%Not measuredOnly GLM 4.6 measured
Neither model has a result on the other 33 boards here
ReasoningLike-for-likePublic boards, shown as published, kept out of the index21
Reasoning benchmarks for GLM 4.6 and Qwen3 235B A22B Thinking 2507
BenchmarkGLM 4.6Qwen3 235B A22B Thinking 2507Leader
GPQA Diamond74.5%Not measuredOnly GLM 4.6 measured
GPQA Diamond (Artificial Analysis run)78.0%Not measuredOnly GLM 4.6 measured
Humanity's Last Exam (Artificial Analysis run)14.5%Not measuredOnly GLM 4.6 measured
GPQA Diamond (Epoch AI run)Not measured80.1%Only Qwen3 235B A22B Thinking 2507 measured
Mystery game puzzlesNot measured9.0%Only Qwen3 235B A22B Thinking 2507 measured
LMArena Text: hard prompts14401416GLM 4.6
LMArena Text: expert14321463Qwen3 235B A22B Thinking 2507
LMArena Text: hard prompts in English14451432GLM 4.6
Kagi LLM Benchmark45.7%Not measuredOnly GLM 4.6 measured
Neither model has a result on the other 14 boards here
MathLike-for-likePublic boards, shown as published, kept out of the index20
Math benchmarks for GLM 4.6 and Qwen3 235B A22B Thinking 2507
BenchmarkGLM 4.6Qwen3 235B A22B Thinking 2507Leader
AIME92.7%Not measuredOnly GLM 4.6 measured
MGSM (Vals AI)89.8%Not measuredOnly GLM 4.6 measured
AA Math Index86.0Not measuredOnly GLM 4.6 measured
AIME 2025 (Artificial Analysis run)86.0%Not measuredOnly GLM 4.6 measured
OTIS Mock AIMENot measured86.7%Only Qwen3 235B A22B Thinking 2507 measured
LMArena Text: math14321412GLM 4.6
LMArena Text: mathematical work14471435GLM 4.6
MathArena16.3%Not measuredOnly GLM 4.6 measured
Neither model has a result on the other 6 boards here
KnowledgeLike-for-likePublic boards, shown as published, kept out of the index31
Knowledge benchmarks for GLM 4.6 and Qwen3 235B A22B Thinking 2507
BenchmarkGLM 4.6Qwen3 235B A22B Thinking 2507Leader
MMLU Pro82.2%Not measuredOnly GLM 4.6 measured
LegalBench (Vals AI)79.6%Not measuredOnly GLM 4.6 measured
CorpFin (Vals AI)56.8%Not measuredOnly GLM 4.6 measured
TaxEval (Vals AI)66.2%Not measuredOnly GLM 4.6 measured
MedQA (Vals AI)92.2%Not measuredOnly GLM 4.6 measured
MMLU Pro (Artificial Analysis run)82.9%Not measuredOnly GLM 4.6 measured
SimpleQA VerifiedNot measured40.4%Only Qwen3 235B A22B Thinking 2507 measured
LMArena Text: science14561433GLM 4.6
LMArena Text: medicine and healthcare14551456Qwen3 235B A22B Thinking 2507
LMArena Text: legal and government14481415GLM 4.6
LMArena Text: business and finance14161412GLM 4.6
Neither model has a result on the other 9 boards here
Web designLike-for-likePublic boards, shown as published, kept out of the index60
Web design benchmarks for GLM 4.6 and Qwen3 235B A22B Thinking 2507
BenchmarkGLM 4.6Qwen3 235B A22B Thinking 2507Leader
Design Arena: websites11741052GLM 4.6
Design Arena: UI components1145942GLM 4.6
Design Arena: game dev1150956GLM 4.6
Design Arena: data viz1161950GLM 4.6
Design Arena: 3D1129998GLM 4.6
Design Arena: all code categories11651032GLM 4.6
Design Arena: SVG1108Not measuredOnly GLM 4.6 measured
Neither model has a result on the other 2 boards here
GeneralLike-for-likePublic boards, shown as published, kept out of the index160
General benchmarks for GLM 4.6 and Qwen3 235B A22B Thinking 2507
BenchmarkGLM 4.6Qwen3 235B A22B Thinking 2507Leader
IFBench43.4%Not measuredOnly GLM 4.6 measured
Long context reasoning (AA-LCR)54.0%Not measuredOnly GLM 4.6 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 Text14401415GLM 4.6
LMArena Text: instruction following14101386GLM 4.6
LMArena Text: creative writing14121384GLM 4.6
LMArena Text: English14491429GLM 4.6
LMArena Text: non-English14261397GLM 4.6
LMArena Text: Chinese14991481GLM 4.6
LMArena Text: German14471408GLM 4.6
LMArena Text: Japanese13931386GLM 4.6
LMArena Text: Korean14001375GLM 4.6
LMArena Text: French1459Not measuredOnly GLM 4.6 measured
LMArena Text: Spanish14361389GLM 4.6
LMArena Text: Russian14191400GLM 4.6
LMArena Text: Polish14471390GLM 4.6
LMArena Text: multi-turn14271406GLM 4.6
LMArena Text: longer queries14221399GLM 4.6
LMArena Text: writing and language14161388GLM 4.6
LMArena Text: entertainment and media14091370GLM 4.6
Neither model has a result on the other 25 boards here
Image generationNot comparablePublic boards, shown as published, kept out of the index0 boards, none shared
Image generation benchmarks for GLM 4.6 and Qwen3 235B A22B Thinking 2507
BenchmarkGLM 4.6Qwen3 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

GLM 4.6$0.0017Fits in one request
Qwen3 235B A22B Thinking 2507$0.0022Fits in one request

GLM 4.6 costs less

Repository review

50K input + 3K output tokens

GLM 4.6$0.037Fits 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

GLM 4.6$0.142Exceeds 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 GLM 4.6 and Qwen3 235B A22B Thinking 2507
SpecificationZ.ai GLM 4.6Qwen3 235B A22B Thinking 2507
MakerZ.aiQwen
WeightsOpen weightsOpen weights
ReleasedSeptember 30, 2025July 25, 2025
Context window205K128K
Input $/Mtok$0.60$0.45
Output $/Mtok$2.20$3.50
Coding index8.8Not benchmarked
Independent measurements behind it20
Cost per solved taskNot computableNot computable
SourceOpenRouter listingOpenRouter listing

Questions

Is GLM 4.6 better than Qwen3 235B A22B Thinking 2507 for coding?

Only GLM 4.6 has a coding score, 8.8 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 work measured by LMArena Text: expert or LMArena Text: medicine and healthcare, where it leads.

Which is cheaper, GLM 4.6 or Qwen3 235B A22B Thinking 2507?

GLM 4.6 costs $0.60 per million input tokens and $2.20 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, GLM 4.6 or Qwen3 235B A22B Thinking 2507?

GLM 4.6, with 205K tokens against 128K.

How many benchmarks is the GLM 4.6 vs Qwen3 235B A22B Thinking 2507 comparison based on?

35 benchmarks have a result for both models as of October 23, 2026. GLM 4.6's coding index rests on 2 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