Qwen3 VL 235B A22B Instruct vs Step 3.5 Flash
Step 3.5 Flash leads Qwen3 VL 235B A22B Instruct on the Artificial Analysis Intelligence Index, 17.0 to 9.9, and Step 3.5 Flash costs 6.3x less per output token ($0.30 vs $1.90 per million). Updated October 21, 2026; 35 benchmarks have a result for both.
Qwen3 VL 235B A22B Instruct or Step 3.5 Flash: which to pick
Neither model has a coding score yet, so the two cannot be ranked on coding.
Pick Qwen3 VL 235B A22B Instruct for work measured by Terminal-Bench Hard (Artificial Analysis run) or AA Intelligence Index, where it leads.
- Shared results
- 35
- Qwen3 VL 235B A22B Instruct only
- 14
- Step 3.5 Flash only
- 0
- 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.
Coding
Vibe Coding Index, six public coding benchmarks
No call
Neither has a coding score yet.
One side not measured
Intelligence
AA Intelligence Index
Step 3.5 Flash
Step 3.5 Flash leads, 17.0 against 9.9.
Same board, both measured
Long documents
Listed context window
No call
Both take 262K tokens.
Listed by OpenRouter
Everyday cost
1K input + 500 output tokens
Step 3.5 Flash
Qwen3 VL 235B A22B Instruct $0.0012, Step 3.5 Flash $0.0003 per turn.
List prices
At a glance
- Coding index
- Qwen3 VL 235B A22B Instruct not benchmarked, Step 3.5 Flash not benchmarked.
- Widest category
- Agentic work, 40.4 points apart on average over 1 shared benchmark.
- Widest single gap
- τ²-Bench Telecom: Qwen3 VL 235B A22B Instruct 54.1% against Step 3.5 Flash 94.4%, 40.4 points apart on a 0 to 100 scale.
- Output price
- Qwen3 VL 235B A22B Instruct $1.90 against Step 3.5 Flash $0.30 per million output tokens, 6.3x apart.
- Context window
- Both take 262K 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.
| Category | Qwen3 VL 235B A22B Instruct | Bars | Step 3.5 Flash | Ties |
|---|---|---|---|---|
| Coding | 2 | 1 | ||
| Intelligence | 0 | 1 | ||
| Agentic work | 0 | 1 | ||
| Reasoning | 3 | 2 | ||
| Math | 1 | 1 | ||
| Knowledge | 4 | 0 | ||
| General | 18 | 1 | ||
| All shared boards | 28 | 7 |
CodingLike-for-likeSix of these boards feed the Vibe Coding Index21
| Benchmark | Qwen3 VL 235B A22B Instruct | Step 3.5 Flash | Leader |
|---|---|---|---|
| LiveCodeBench (Artificial Analysis run) | 64.6% | Not measured | Only Qwen3 VL 235B A22B Instruct measured |
| Terminal-Bench Hard (Artificial Analysis run) | 11.4% | 32.6% | Step 3.5 Flash |
| LMArena Text: coding | 1438 | 1436 | Qwen3 VL 235B A22B Instruct |
| LMArena Text: software and IT | 1442 | 1434 | Qwen3 VL 235B A22B Instruct |
| Neither model has a result on the other 35 boards here | |||
IntelligenceLike-for-likePublic boards, shown as published, kept out of the index01
| Benchmark | Qwen3 VL 235B A22B Instruct | Step 3.5 Flash | Leader |
|---|---|---|---|
| AA Intelligence Index | 9.9 | 17.0 | Step 3.5 Flash |
| Neither model has a result on the other 1 board here | |||
Agentic workLike-for-likePublic boards, shown as published, kept out of the index01
| Benchmark | Qwen3 VL 235B A22B Instruct | Step 3.5 Flash | Leader |
|---|---|---|---|
| τ²-Bench Telecom | 54.1% | 94.4% | Step 3.5 Flash |
| 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 VL 235B A22B Instruct | Step 3.5 Flash | Leader |
|---|---|---|---|
| GPQA Diamond (Artificial Analysis run) | 77.2% | 83.1% | Step 3.5 Flash |
| Humanity's Last Exam (Artificial Analysis run) | 11.9% | 24.5% | Step 3.5 Flash |
| LMArena Text: hard prompts | 1428 | 1411 | Qwen3 VL 235B A22B Instruct |
| LMArena Text: expert | 1442 | 1421 | Qwen3 VL 235B A22B Instruct |
| LMArena Text: hard prompts in English | 1439 | 1421 | Qwen3 VL 235B A22B Instruct |
| LMArena Vision: diagrams | 1255 | Not measured | Only Qwen3 VL 235B A22B Instruct measured |
| LMArena Vision: homework | 1277 | Not measured | Only Qwen3 VL 235B A22B Instruct measured |
| Neither model has a result on the other 15 boards here | |||
MathLike-for-likePublic boards, shown as published, kept out of the index11
| Benchmark | Qwen3 VL 235B A22B Instruct | Step 3.5 Flash | Leader |
|---|---|---|---|
| AA Math Index | 88.3 | Not measured | Only Qwen3 VL 235B A22B Instruct measured |
| AIME 2025 (Artificial Analysis run) | 88.3% | Not measured | Only Qwen3 VL 235B A22B Instruct measured |
| LMArena Text: math | 1425 | 1408 | Qwen3 VL 235B A22B Instruct |
| LMArena Text: mathematical work | 1419 | 1421 | Step 3.5 Flash |
| Neither model has a result on the other 9 boards here | |||
KnowledgeLike-for-likePublic boards, shown as published, kept out of the index40
| Benchmark | Qwen3 VL 235B A22B Instruct | Step 3.5 Flash | Leader |
|---|---|---|---|
| MMLU Pro (Artificial Analysis run) | 83.6% | Not measured | Only Qwen3 VL 235B A22B Instruct measured |
| LMArena Text: science | 1444 | 1415 | Qwen3 VL 235B A22B Instruct |
| LMArena Text: medicine and healthcare | 1431 | 1408 | Qwen3 VL 235B A22B Instruct |
| LMArena Text: legal and government | 1438 | 1408 | Qwen3 VL 235B A22B Instruct |
| LMArena Text: business and finance | 1433 | 1394 | Qwen3 VL 235B A22B Instruct |
| Neither model has a result on the other 13 boards here | |||
Web designNot comparablePublic boards, shown as published, kept out of the index0 boards, none shared
| Benchmark | Qwen3 VL 235B A22B Instruct | Step 3.5 Flash | Leader |
|---|---|---|---|
| Neither model has a result on 9 boards here | |||
GeneralLike-for-likePublic boards, shown as published, kept out of the index181
| Benchmark | Qwen3 VL 235B A22B Instruct | Step 3.5 Flash | Leader |
|---|---|---|---|
| IFBench | 56.5% | 66.5% | Step 3.5 Flash |
| Long context reasoning (AA-LCR) | 63.7% | 63.0% | Qwen3 VL 235B A22B Instruct |
| LMArena Text | 1420 | 1403 | Qwen3 VL 235B A22B Instruct |
| LMArena Text: instruction following | 1406 | 1385 | Qwen3 VL 235B A22B Instruct |
| LMArena Text: creative writing | 1366 | 1357 | Qwen3 VL 235B A22B Instruct |
| LMArena Vision | 1247 | Not measured | Only Qwen3 VL 235B A22B Instruct measured |
| LMArena Text: English | 1429 | 1417 | Qwen3 VL 235B A22B Instruct |
| LMArena Text: non-English | 1405 | 1385 | Qwen3 VL 235B A22B Instruct |
| LMArena Text: Chinese | 1463 | 1446 | Qwen3 VL 235B A22B Instruct |
| LMArena Text: German | 1424 | 1405 | Qwen3 VL 235B A22B Instruct |
| LMArena Text: Japanese | 1385 | 1353 | Qwen3 VL 235B A22B Instruct |
| LMArena Text: Korean | 1394 | 1351 | Qwen3 VL 235B A22B Instruct |
| LMArena Text: French | 1452 | 1421 | Qwen3 VL 235B A22B Instruct |
| LMArena Text: Spanish | 1428 | 1419 | Qwen3 VL 235B A22B Instruct |
| LMArena Text: Russian | 1408 | 1385 | Qwen3 VL 235B A22B Instruct |
| LMArena Text: Polish | 1419 | 1389 | Qwen3 VL 235B A22B Instruct |
| LMArena Text: multi-turn | 1428 | 1404 | Qwen3 VL 235B A22B Instruct |
| LMArena Text: longer queries | 1419 | 1402 | Qwen3 VL 235B A22B Instruct |
| LMArena Text: writing and language | 1392 | 1372 | Qwen3 VL 235B A22B Instruct |
| LMArena Text: entertainment and media | 1368 | 1360 | Qwen3 VL 235B A22B Instruct |
| LMArena Vision: English | 1256 | Not measured | Only Qwen3 VL 235B A22B Instruct measured |
| LMArena Vision: Chinese | 1283 | Not measured | Only Qwen3 VL 235B A22B Instruct measured |
| LMArena Vision: reading text | 1253 | Not measured | Only Qwen3 VL 235B A22B Instruct measured |
| LMArena Vision: humor | 1209 | Not measured | Only Qwen3 VL 235B A22B Instruct measured |
| LMArena Vision: captioning | 1281 | Not measured | Only Qwen3 VL 235B A22B Instruct measured |
| LMArena Vision: recognizing things | 1249 | Not measured | Only Qwen3 VL 235B A22B Instruct measured |
| LMArena Vision: creative writing | 1220 | Not measured | Only Qwen3 VL 235B A22B Instruct measured |
| Neither model has a result on the other 19 boards here | |||
Image generationNot comparablePublic boards, shown as published, kept out of the index0 boards, none shared
| Benchmark | Qwen3 VL 235B A22B Instruct | Step 3.5 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
Step 3.5 Flash costs less
Repository review
50K input + 3K output tokens
Step 3.5 Flash costs less
Agent loop
200K input + 10K output tokens
Step 3.5 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 VL 235B A22B Instruct | StepFun Step 3.5 Flash |
|---|---|---|
| Maker | Qwen | StepFun |
| Weights | Open weights | Open weights |
| Released | September 23, 2025 | January 29, 2026 |
| Context window | 262K | 262K |
| Input $/Mtok | $0.21 | $0.10 |
| Output $/Mtok | $1.90 | $0.30 |
| Coding index | Not benchmarked | Not benchmarked |
| Independent measurements behind it | 0 | 0 |
| Cost per solved task | Not computable | Not computable |
| Source | OpenRouter listing | OpenRouter listing |
Questions
Is Qwen3 VL 235B A22B Instruct better than Step 3.5 Flash for coding?
Neither model has a coding score yet, so the two cannot be ranked on coding. Pick Qwen3 VL 235B A22B Instruct for work measured by Terminal-Bench Hard (Artificial Analysis run) or AA Intelligence Index, where it leads.
Which is cheaper, Qwen3 VL 235B A22B Instruct or Step 3.5 Flash?
Qwen3 VL 235B A22B Instruct costs $0.21 per million input tokens and $1.90 per million output tokens; Step 3.5 Flash costs $0.10 and $0.30. Cost per solved task is known for neither of them.
Which has the bigger context window, Qwen3 VL 235B A22B Instruct or Step 3.5 Flash?
Neither: both take 262K tokens.
How many benchmarks is the Qwen3 VL 235B A22B Instruct vs Step 3.5 Flash comparison based on?
35 benchmarks have a result for both models as of October 21, 2026. Qwen3 VL 235B A22B Instruct's coding index rests on 0 independent benchmarks and Step 3.5 Flash's on 0. Where only one model has a result, the page says so instead of naming a winner.