Qwen3 235B A22B Thinking 2507 vs Qwen3.5-Flash
Qwen3.5-Flash costs 13.5x less per output token than Qwen3 235B A22B Thinking 2507 ($0.26 vs $3.50 per million). Updated October 23, 2026; 33 benchmarks have a result for both.
Qwen3 235B A22B Thinking 2507 or Qwen3.5-Flash: which to pick
Only Qwen3.5-Flash has a coding score, 41.1 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: coding or LMArena Text: software and IT, where it leads.
- Shared results
- 33
- Qwen3 235B A22B Thinking 2507 only
- 13
- Qwen3.5-Flash only
- 22
- Like-for-like categories
- 5 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.
Coding
Vibe Coding Index, six public coding benchmarks
No call
Only Qwen3.5-Flash has a coding score, so no call is made.
One side not measured
Intelligence
AA Intelligence Index
No call
Only Qwen3 235B A22B Thinking 2507 has a AA Intelligence Index result, 12.7.
Directional only
Agentic work
AA Agentic Index
No call
Only Qwen3 235B A22B Thinking 2507 has a AA Agentic Index result, 1.3.
Directional only
Reasoning
GPQA Diamond
No call
Only Qwen3.5-Flash has a GPQA Diamond result, 82.8%.
Directional only
Math
AIME
No call
Only Qwen3.5-Flash has a AIME result, 92.5%.
Directional only
Knowledge
MMLU Pro
No call
Only Qwen3.5-Flash has a MMLU Pro result, 84.1%.
Directional only
Web design
Design Arena: websites
No call
Only Qwen3 235B A22B Thinking 2507 has a Design Arena: websites result, 1052.
Directional only
Long documents
Listed context window
Qwen3.5-Flash
Qwen3.5-Flash takes 1M tokens against 128K.
Listed by OpenRouter
Everyday cost
1K input + 500 output tokens
Qwen3.5-Flash
Qwen3 235B A22B Thinking 2507 $0.0022, Qwen3.5-Flash $0.0002 per turn.
List prices
Cost per solved task
One SWE-bench Verified task, solved
No call
Only Qwen3.5-Flash has the number, $0.40.
Directional only
At a glance
- Coding index
- Qwen3 235B A22B Thinking 2507 not benchmarked, Qwen3.5-Flash 41.1 from 2 benchmarks.
- Widest category
- Knowledge, 7.4 points apart on average over 5 shared benchmarks.
- Widest single gap
- SimpleQA Verified: Qwen3 235B A22B Thinking 2507 40.4% against Qwen3.5-Flash 20.3%, 20.1 points apart on a 0 to 100 scale.
- Output price
- Qwen3 235B A22B Thinking 2507 $3.50 against Qwen3.5-Flash $0.26 per million output tokens, 13.5x apart.
- Context window
- Qwen3 235B A22B Thinking 2507 takes 128K tokens, Qwen3.5-Flash 1M.
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.
| Category | Qwen3 235B A22B Thinking 2507 | Bars | Qwen3.5-Flash | Ties |
|---|---|---|---|---|
| Coding | 2 | 0 | ||
| Reasoning | 3 | 2 | ||
| Math | 3 | 0 | ||
| Knowledge | 5 | 0 | ||
| General | 16 | 2 | ||
| All shared boards | 29 | 4 |
CodingLike-for-likeSix of these boards feed the Vibe Coding Index20
| Benchmark | Qwen3 235B A22B Thinking 2507 | Qwen3.5-Flash | Leader |
|---|---|---|---|
| LMArena WebDev | Not measured | 1244 | Only Qwen3.5-Flash measured |
| AA Coding Index | 22.1 | Not measured | Only Qwen3 235B A22B Thinking 2507 measured |
| LiveCodeBench (Vals AI) | Not measured | 83.3% | Only Qwen3.5-Flash measured |
| SWE-bench Verified (Vals AI) | Not measured | 64.4% | Only Qwen3.5-Flash measured |
| WeirdML | 41.0% | Not measured | Only Qwen3 235B A22B Thinking 2507 measured |
| ALE-Bench | Not measured | 265.9 | Only Qwen3.5-Flash measured |
| LMArena Text: coding | 1424 | 1412 | Qwen3 235B A22B Thinking 2507 |
| LMArena Text: software and IT | 1431 | 1416 | Qwen3 235B A22B Thinking 2507 |
| LMArena WebDev: React | Not measured | 1229 | Only Qwen3.5-Flash measured |
| LMArena WebDev: HTML | Not measured | 1291 | Only Qwen3.5-Flash measured |
| Neither model has a result on the other 33 boards here | |||
IntelligenceDirectional onlyPublic boards, shown as published, kept out of the index2 boards, none shared
| Benchmark | Qwen3 235B A22B Thinking 2507 | Qwen3.5-Flash | Leader |
|---|---|---|---|
| AA Intelligence Index | 12.7 | Not measured | Only Qwen3 235B A22B Thinking 2507 measured |
| Epoch Capabilities Index | Not measured | 143.5 | Only Qwen3.5-Flash measured |
Agentic workDirectional onlyPublic boards, shown as published, kept out of the index2 boards, none shared
| Benchmark | Qwen3 235B A22B Thinking 2507 | Qwen3.5-Flash | Leader |
|---|---|---|---|
| AA Agentic Index | 1.3 | Not measured | Only Qwen3 235B A22B Thinking 2507 measured |
| Vending-Bench 2 | -$11 | Not measured | Only Qwen3 235B A22B Thinking 2507 measured |
| Neither model has a result on the other 40 boards here | |||
ReasoningLike-for-likePublic boards, shown as published, kept out of the index32
| Benchmark | Qwen3 235B A22B Thinking 2507 | Qwen3.5-Flash | Leader |
|---|---|---|---|
| GPQA Diamond | Not measured | 82.8% | Only Qwen3.5-Flash measured |
| MMMU Pro (Vals AI) | Not measured | 81.9% | Only Qwen3.5-Flash measured |
| GPQA Diamond (Epoch AI run) | 80.1% | 82.3% | Qwen3.5-Flash |
| Chess puzzles | Not measured | 14.0% | Only Qwen3.5-Flash measured |
| Mystery game puzzles | 9.0% | 20.0% | Qwen3.5-Flash |
| LMArena Text: hard prompts | 1416 | 1403 | Qwen3 235B A22B Thinking 2507 |
| LMArena Text: expert | 1463 | 1407 | Qwen3 235B A22B Thinking 2507 |
| LMArena Text: hard prompts in English | 1432 | 1409 | Qwen3 235B A22B Thinking 2507 |
| Neither model has a result on the other 15 boards here | |||
MathLike-for-likePublic boards, shown as published, kept out of the index30
| Benchmark | Qwen3 235B A22B Thinking 2507 | Qwen3.5-Flash | Leader |
|---|---|---|---|
| AIME | Not measured | 92.5% | Only Qwen3.5-Flash measured |
| OTIS Mock AIME | 86.7% | 84.4% | Qwen3 235B A22B Thinking 2507 |
| FrontierMath, tiers 1 to 3 | Not measured | 18.2% | Only Qwen3.5-Flash measured |
| LMArena Text: math | 1412 | 1407 | Qwen3 235B A22B Thinking 2507 |
| LMArena Text: mathematical work | 1435 | 1413 | Qwen3 235B A22B Thinking 2507 |
| Neither model has a result on the other 9 boards here | |||
KnowledgeLike-for-likePublic boards, shown as published, kept out of the index50
| Benchmark | Qwen3 235B A22B Thinking 2507 | Qwen3.5-Flash | Leader |
|---|---|---|---|
| MMLU Pro | Not measured | 84.1% | Only Qwen3.5-Flash measured |
| LegalBench (Vals AI) | Not measured | 84.3% | Only Qwen3.5-Flash measured |
| CaseLaw (Vals AI) | Not measured | 56.0% | Only Qwen3.5-Flash measured |
| CorpFin (Vals AI) | Not measured | 63.6% | Only Qwen3.5-Flash measured |
| TaxEval (Vals AI) | Not measured | 72.2% | Only Qwen3.5-Flash measured |
| MortgageTax (Vals AI) | Not measured | 67.4% | Only Qwen3.5-Flash measured |
| MedCode (Vals AI) | Not measured | 33.0% | Only Qwen3.5-Flash measured |
| SimpleQA Verified | 40.4% | 20.3% | Qwen3 235B A22B Thinking 2507 |
| LMArena Text: science | 1433 | 1415 | Qwen3 235B A22B Thinking 2507 |
| LMArena Text: medicine and healthcare | 1456 | 1407 | Qwen3 235B A22B Thinking 2507 |
| LMArena Text: legal and government | 1415 | 1397 | Qwen3 235B A22B Thinking 2507 |
| LMArena Text: business and finance | 1412 | 1394 | Qwen3 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
| Benchmark | Qwen3 235B A22B Thinking 2507 | Qwen3.5-Flash | Leader |
|---|---|---|---|
| Design Arena: websites | 1052 | Not measured | Only Qwen3 235B A22B Thinking 2507 measured |
| Design Arena: UI components | 942 | Not measured | Only Qwen3 235B A22B Thinking 2507 measured |
| Design Arena: game dev | 956 | Not measured | Only Qwen3 235B A22B Thinking 2507 measured |
| Design Arena: data viz | 950 | Not measured | Only Qwen3 235B A22B Thinking 2507 measured |
| Design Arena: 3D | 998 | Not measured | Only Qwen3 235B A22B Thinking 2507 measured |
| Design Arena: all code categories | 1032 | Not measured | Only 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 index162
| Benchmark | Qwen3 235B A22B Thinking 2507 | Qwen3.5-Flash | Leader |
|---|---|---|---|
| SAGE (Vals AI) | Not measured | 42.5% | Only Qwen3.5-Flash measured |
| MedScribe (Vals AI) | Not measured | 70.6% | Only Qwen3.5-Flash measured |
| DTBench | 80.3% | 82.9% | Qwen3.5-Flash |
| LMCA | 29.3% | 29.1% | Qwen3 235B A22B Thinking 2507 |
| Fiction.LiveBench | 75.0% | Not measured | Only Qwen3 235B A22B Thinking 2507 measured |
| Creative writing (Lech Mazur) | 82.4% | Not measured | Only Qwen3 235B A22B Thinking 2507 measured |
| LMArena Text | 1415 | 1397 | Qwen3 235B A22B Thinking 2507 |
| LMArena Text: instruction following | 1386 | 1375 | Qwen3 235B A22B Thinking 2507 |
| LMArena Text: creative writing | 1384 | 1343 | Qwen3 235B A22B Thinking 2507 |
| LMArena Text: English | 1429 | 1404 | Qwen3 235B A22B Thinking 2507 |
| LMArena Text: non-English | 1397 | 1385 | Qwen3 235B A22B Thinking 2507 |
| LMArena Text: Chinese | 1481 | 1446 | Qwen3 235B A22B Thinking 2507 |
| LMArena Text: German | 1408 | 1391 | Qwen3 235B A22B Thinking 2507 |
| LMArena Text: Japanese | 1386 | 1369 | Qwen3 235B A22B Thinking 2507 |
| LMArena Text: Korean | 1375 | 1344 | Qwen3 235B A22B Thinking 2507 |
| LMArena Text: French | Not measured | 1413 | Only Qwen3.5-Flash measured |
| LMArena Text: Spanish | 1389 | 1399 | Qwen3.5-Flash |
| LMArena Text: Russian | 1400 | 1380 | Qwen3 235B A22B Thinking 2507 |
| LMArena Text: Polish | 1390 | 1376 | Qwen3 235B A22B Thinking 2507 |
| LMArena Text: multi-turn | 1406 | 1393 | Qwen3 235B A22B Thinking 2507 |
| LMArena Text: longer queries | 1399 | 1392 | Qwen3 235B A22B Thinking 2507 |
| LMArena Text: writing and language | 1388 | 1369 | Qwen3 235B A22B Thinking 2507 |
| LMArena Text: entertainment and media | 1370 | 1347 | Qwen3 235B A22B Thinking 2507 |
| 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
| Benchmark | Qwen3 235B A22B Thinking 2507 | Qwen3.5-Flash | Leader |
|---|---|---|---|
| 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
Qwen3.5-Flash costs less
Repository review
50K input + 3K output tokens
Qwen3.5-Flash costs less
Agent loop
200K input + 10K output tokens
Qwen3.5-Flash 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.
| Specification | Qwen3 235B A22B Thinking 2507 | Qwen3.5-Flash |
|---|---|---|
| Maker | Qwen | Qwen |
| Weights | Open weights | No public weights listed |
| Released | July 25, 2025 | February 25, 2026 |
| Context window | 128K | 1M |
| Input $/Mtok | $0.45 | $0.065 |
| Output $/Mtok | $3.50 | $0.26 |
| Coding index | Not benchmarked | 41.1 |
| Independent measurements behind it | 0 | 2 |
| Cost per solved task | Not computable | $0.40 |
| Source | OpenRouter listing | OpenRouter listing |
Questions
Is Qwen3 235B A22B Thinking 2507 better than Qwen3.5-Flash for coding?
Only Qwen3.5-Flash has a coding score, 41.1 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: coding or LMArena Text: software and IT, where it leads.
Which is cheaper, Qwen3 235B A22B Thinking 2507 or Qwen3.5-Flash?
Qwen3 235B A22B Thinking 2507 costs $0.45 per million input tokens and $3.50 per million output tokens; Qwen3.5-Flash costs $0.065 and $0.26. Cost per solved task is known for Qwen3.5-Flash only.
Which has the bigger context window, Qwen3 235B A22B Thinking 2507 or Qwen3.5-Flash?
Qwen3.5-Flash, with 1M tokens against 128K.
How many benchmarks is the Qwen3 235B A22B Thinking 2507 vs Qwen3.5-Flash comparison based on?
33 benchmarks have a result for both models as of October 23, 2026. Qwen3 235B A22B Thinking 2507's coding index rests on 0 independent benchmarks and Qwen3.5-Flash's on 2. Where only one model has a result, the page says so instead of naming a winner.