Qwen3.5-Flash vs Extract
Updated October 22, 2026; no benchmark has tested both models yet.
- Coding
- 41.1
- Overall
- 50.1
- Output $/M
- $0.26
- Coding
- No data
- Overall
- No data
- Output $/M
- No data
Qwen3.5-Flash or Extract: which to pick
Only Qwen3.5-Flash has a coding score, 41.1 on the Vibe Coding Index; Extract has none, so the two cannot be ranked on coding.
Compare them on price, context and the boards below.
No benchmark has tested both yet. Qwen3.5-Flash has results on LMArena WebDev, LiveCodeBench (Vals AI) and SWE-bench Verified (Vals AI); Extract has none yet. Compare them on price and context instead.
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
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
Long documents
Listed context window
No call
Both take 1M tokens.
Listed by OpenRouter
Everyday cost
1K input + 500 output tokens
Extract
Qwen3.5-Flash $0.0002, Extract $0.0000 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
- Qwen3.5-Flash 41.1 from 2 benchmarks, Extract not benchmarked.
- Widest category
- None yet: no category has been tested on both models.
- Widest single gap
- None yet: no benchmark has tested both models.
- Output price
- Qwen3.5-Flash $0.26 against Extract Free per million output tokens.
- Context window
- Both take 1M tokens.
Benchmark by benchmark
Every board either model has a result on, a category to a fold; bars are each result on a 0 to 100 scale. Only one model is measured on each, so no row names a leader.
CodingDirectional onlySix of these boards feed the Vibe Coding Index8 boards measured
| Benchmark | Qwen3.5-Flash | Extract | Leader |
|---|---|---|---|
| LMArena WebDev | 1244 | Not measured | Only Qwen3.5-Flash measured |
| LiveCodeBench (Vals AI) | 83.3% | Not measured | Only Qwen3.5-Flash measured |
| SWE-bench Verified (Vals AI) | 64.4% | Not measured | Only Qwen3.5-Flash measured |
| ALE-Bench | 265.9 | Not measured | Only Qwen3.5-Flash measured |
| LMArena Text: coding | 1412 | Not measured | Only Qwen3.5-Flash measured |
| LMArena Text: software and IT | 1416 | Not measured | Only Qwen3.5-Flash measured |
| LMArena WebDev: React | 1229 | Not measured | Only Qwen3.5-Flash measured |
| LMArena WebDev: HTML | 1291 | Not measured | Only Qwen3.5-Flash measured |
| Neither model has a result on the other 31 boards here | |||
IntelligenceDirectional onlyPublic boards, shown as published, kept out of the index1 board measured
| Benchmark | Qwen3.5-Flash | Extract | Leader |
|---|---|---|---|
| Epoch Capabilities Index | 143.5 | Not measured | Only Qwen3.5-Flash measured |
| Neither model has a result on the other 1 board here | |||
ReasoningDirectional onlyPublic boards, shown as published, kept out of the index8 boards measured
| Benchmark | Qwen3.5-Flash | Extract | Leader |
|---|---|---|---|
| GPQA Diamond | 82.8% | Not measured | Only Qwen3.5-Flash measured |
| MMMU Pro (Vals AI) | 81.9% | Not measured | Only Qwen3.5-Flash measured |
| GPQA Diamond (Epoch AI run) | 82.3% | Not measured | Only Qwen3.5-Flash measured |
| Chess puzzles | 14.0% | Not measured | Only Qwen3.5-Flash measured |
| Mystery game puzzles | 20.0% | Not measured | Only Qwen3.5-Flash measured |
| LMArena Text: hard prompts | 1403 | Not measured | Only Qwen3.5-Flash measured |
| LMArena Text: expert | 1407 | Not measured | Only Qwen3.5-Flash measured |
| LMArena Text: hard prompts in English | 1409 | Not measured | Only Qwen3.5-Flash measured |
| Neither model has a result on the other 14 boards here | |||
MathDirectional onlyPublic boards, shown as published, kept out of the index5 boards measured
| Benchmark | Qwen3.5-Flash | Extract | Leader |
|---|---|---|---|
| AIME | 92.5% | Not measured | Only Qwen3.5-Flash measured |
| OTIS Mock AIME | 84.4% | Not measured | Only Qwen3.5-Flash measured |
| FrontierMath, tiers 1 to 3 | 18.2% | Not measured | Only Qwen3.5-Flash measured |
| LMArena Text: math | 1407 | Not measured | Only Qwen3.5-Flash measured |
| LMArena Text: mathematical work | 1413 | Not measured | Only Qwen3.5-Flash measured |
| Neither model has a result on the other 8 boards here | |||
KnowledgeDirectional onlyPublic boards, shown as published, kept out of the index12 boards measured
| Benchmark | Qwen3.5-Flash | Extract | Leader |
|---|---|---|---|
| MMLU Pro | 84.1% | Not measured | Only Qwen3.5-Flash measured |
| LegalBench (Vals AI) | 84.3% | Not measured | Only Qwen3.5-Flash measured |
| CaseLaw (Vals AI) | 56.0% | Not measured | Only Qwen3.5-Flash measured |
| CorpFin (Vals AI) | 63.6% | Not measured | Only Qwen3.5-Flash measured |
| TaxEval (Vals AI) | 72.2% | Not measured | Only Qwen3.5-Flash measured |
| MortgageTax (Vals AI) | 67.4% | Not measured | Only Qwen3.5-Flash measured |
| MedCode (Vals AI) | 33.0% | Not measured | Only Qwen3.5-Flash measured |
| SimpleQA Verified | 20.3% | Not measured | Only Qwen3.5-Flash measured |
| LMArena Text: science | 1415 | Not measured | Only Qwen3.5-Flash measured |
| LMArena Text: medicine and healthcare | 1407 | Not measured | Only Qwen3.5-Flash measured |
| LMArena Text: legal and government | 1398 | Not measured | Only Qwen3.5-Flash measured |
| LMArena Text: business and finance | 1394 | Not measured | Only Qwen3.5-Flash measured |
| Neither model has a result on the other 6 boards here | |||
GeneralDirectional onlyPublic boards, shown as published, kept out of the index21 boards measured
| Benchmark | Qwen3.5-Flash | Extract | Leader |
|---|---|---|---|
| SAGE (Vals AI) | 42.5% | Not measured | Only Qwen3.5-Flash measured |
| MedScribe (Vals AI) | 70.6% | Not measured | Only Qwen3.5-Flash measured |
| DTBench | 82.9% | Not measured | Only Qwen3.5-Flash measured |
| LMCA | 29.1% | Not measured | Only Qwen3.5-Flash measured |
| LMArena Text | 1397 | Not measured | Only Qwen3.5-Flash measured |
| LMArena Text: instruction following | 1374 | Not measured | Only Qwen3.5-Flash measured |
| LMArena Text: creative writing | 1343 | Not measured | Only Qwen3.5-Flash measured |
| LMArena Text: English | 1404 | Not measured | Only Qwen3.5-Flash measured |
| LMArena Text: non-English | 1385 | Not measured | Only Qwen3.5-Flash measured |
| LMArena Text: Chinese | 1445 | Not measured | Only Qwen3.5-Flash measured |
| LMArena Text: German | 1390 | Not measured | Only Qwen3.5-Flash measured |
| LMArena Text: Japanese | 1369 | Not measured | Only Qwen3.5-Flash measured |
| LMArena Text: Korean | 1344 | Not measured | Only Qwen3.5-Flash measured |
| LMArena Text: French | 1413 | Not measured | Only Qwen3.5-Flash measured |
| LMArena Text: Spanish | 1400 | Not measured | Only Qwen3.5-Flash measured |
| LMArena Text: Russian | 1380 | Not measured | Only Qwen3.5-Flash measured |
| LMArena Text: Polish | 1376 | Not measured | Only Qwen3.5-Flash measured |
| LMArena Text: multi-turn | 1393 | Not measured | Only Qwen3.5-Flash measured |
| LMArena Text: longer queries | 1392 | Not measured | Only Qwen3.5-Flash measured |
| LMArena Text: writing and language | 1369 | Not measured | Only Qwen3.5-Flash measured |
| LMArena Text: entertainment and media | 1347 | Not measured | Only Qwen3.5-Flash measured |
| Neither model has a result on the other 25 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
Extract costs less
Repository review
50K input + 3K output tokens
Extract costs less
Agent loop
200K input + 10K output tokens
Extract 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 | Qwen3.5-Flash | ScaleDown Extract |
|---|---|---|
| Maker | Qwen | Scaledown |
| Weights | No public weights listed | No public weights listed |
| Released | February 25, 2026 | October 13, 2026 |
| Context window | 1M | 1M |
| Input $/Mtok | $0.065 | $0.04 |
| Output $/Mtok | $0.26 | Free |
| Coding index | 41.1 | Not benchmarked |
| Independent measurements behind it | 2 | 0 |
| Cost per solved task | $0.40 | Not computable |
| Source | OpenRouter listing | OpenRouter listing |
Questions
Is Qwen3.5-Flash better than Extract for coding?
Only Qwen3.5-Flash has a coding score, 41.1 on the Vibe Coding Index; Extract has none, so the two cannot be ranked on coding. Compare them on price, context and the boards below.
Which is cheaper, Qwen3.5-Flash or Extract?
Qwen3.5-Flash costs $0.065 per million input tokens and $0.26 per million output tokens; Extract costs $0.04 and Free. Cost per solved task is known for Qwen3.5-Flash only.
Which has the bigger context window, Qwen3.5-Flash or Extract?
Neither: both take 1M tokens.
Has any benchmark tested both Qwen3.5-Flash and Extract?
No benchmark has tested both yet. Qwen3.5-Flash has results on LMArena WebDev, LiveCodeBench (Vals AI) and SWE-bench Verified (Vals AI); Extract has none yet. Compare them on price and context instead. Data as of October 22, 2026.