GPT-5.2 vs Qwen3.8 27B
Qwen3.8 27B leads GPT-5.2 on coding, 74.2 to 55.9 on the Vibe Coding Index, and Qwen3.8 27B costs 5.5x less per output token ($2.55 vs $14.00 per million). Updated October 22, 2026; 60 benchmarks have a result for both.
GPT-5.2 or Qwen3.8 27B: which to pick
Pick Qwen3.8 27B for coding: it scores 74.2 on the Vibe Coding Index against 55.9 for GPT-5.2.
Pick GPT-5.2 for work measured by LiveCodeBench (Vals AI) or Epoch Capabilities Index, where it leads.
- Shared results
- 60
- GPT-5.2 only
- 54
- Qwen3.8 27B only
- 25
- Like-for-like categories
- 6 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
Qwen3.8 27B
Qwen3.8 27B scores 74.2 against 55.9.
3 and 4 benchmarks behind the scores
Intelligence
AA Intelligence Index
Qwen3.8 27B
Qwen3.8 27B leads, 33.7 against 30.4.
Same board, both measured
Agentic work
AA Agentic Index
No call
Only Qwen3.8 27B has a AA Agentic Index result, 45.8.
Directional only
Reasoning
GPQA Diamond
GPT-5.2
GPT-5.2 leads, 91.7% against 88.9%.
Same board, both measured
Math
AIME
No call
Only GPT-5.2 has a AIME result, 96.9%.
Directional only
Knowledge
MMLU Pro
GPT-5.2
GPT-5.2 leads, 86.2% against 84.3%.
Same board, both measured
Web design
Design Arena: websites
No call
Only GPT-5.2 has a Design Arena: websites result, 1195.
Directional only
Long documents
Listed context window
Qwen3.8 27B
Qwen3.8 27B takes 1M tokens against 400K.
Listed by OpenRouter
Everyday cost
1K input + 500 output tokens
Qwen3.8 27B
GPT-5.2 $0.0088, Qwen3.8 27B $0.0017 per turn.
List prices
Cost per solved task
One SWE-bench Verified task, solved
Qwen3.8 27B
GPT-5.2 $1.91, Qwen3.8 27B $1.26.
Published attempt cost and pass rate
At a glance
- Coding index
- GPT-5.2 55.9 from 3 benchmarks, Qwen3.8 27B 74.2 from 4 benchmarks.
- Widest category
- Coding, 12.3 points apart on average over 7 shared benchmarks.
- Widest single gap
- LMArena WebDev: GPT-5.2 1417 against Qwen3.8 27B 1593, 25.1 points apart on a 0 to 100 scale.
- Output price
- GPT-5.2 $14.00 against Qwen3.8 27B $2.55 per million output tokens, 5.5x apart.
- Context window
- GPT-5.2 takes 400K tokens, Qwen3.8 27B 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 | GPT-5.2 | Bars | Qwen3.8 27B | Ties |
|---|---|---|---|---|
| Coding | 1 | 6 | ||
| Intelligence | 1 | 1 | ||
| Reasoning | 4 | 7 | ||
| Math | 0 | 3 | ||
| Knowledge | 5 | 4 | ||
| General | 6 | 22 | ||
| All shared boards | 17 | 43 |
CodingLike-for-likeSix of these boards feed the Vibe Coding Index16
| Benchmark | GPT-5.2 | Qwen3.8 27B | Leader |
|---|---|---|---|
| SWE-bench Verified | 72.8% | Not measured | Only GPT-5.2 measured |
| SWE-bench Multilingual | 66.7% | Not measured | Only GPT-5.2 measured |
| LiveBench coding | Not measured | 86.8% | Only Qwen3.8 27B measured |
| LMArena WebDev | 1417 | 1593 | Qwen3.8 27B |
| AA Coding Index | Not measured | 68.1 | Only Qwen3.8 27B measured |
| LiveCodeBench (Vals AI) | 85.4% | 84.0% | GPT-5.2 |
| SWE-bench Verified (Vals AI) | 75.8% | 86.0% | Qwen3.8 27B |
| Vibe Code Bench (Vals AI) | 53.5% | 64.8% | Qwen3.8 27B |
| Code migration (Vals AI) | Not measured | 14.2% | Only Qwen3.8 27B measured |
| ProgramBench (Vals AI) | Not measured | 0.0% | Only Qwen3.8 27B measured |
| IOI (Vals AI) | Not measured | 39.1% | Only Qwen3.8 27B measured |
| Terminal-Bench 2.1 (Vals AI) | Not measured | 58.4% | Only Qwen3.8 27B measured |
| SkillsBench (Vals AI) | Not measured | 38.1% | Only Qwen3.8 27B measured |
| CyberBench (Vals AI) | Not measured | 67.8% | Only Qwen3.8 27B measured |
| LiveCodeBench (Artificial Analysis run) | 89.4% | Not measured | Only GPT-5.2 measured |
| SciCode | Not measured | 46.6% | Only Qwen3.8 27B measured |
| Terminal-Bench Hard (Artificial Analysis run) | 47.0% | Not measured | Only GPT-5.2 measured |
| Terminal-Bench | 64.9% | Not measured | Only GPT-5.2 measured |
| GSO-Bench | 27.5% | Not measured | Only GPT-5.2 measured |
| WeirdML | 72.2% | Not measured | Only GPT-5.2 measured |
| ALE-Bench | 1293.5 | Not measured | Only GPT-5.2 measured |
| LMArena Text: coding | 1442 | 1483 | Qwen3.8 27B |
| LMArena Text: software and IT | 1434 | 1478 | Qwen3.8 27B |
| LMArena WebDev: React | Not measured | 1602 | Only Qwen3.8 27B measured |
| LMArena WebDev: HTML | 1427 | 1570 | Qwen3.8 27B |
| LMArena WebDev: from an image | Not measured | 1577 | Only Qwen3.8 27B measured |
| Neither model has a result on the other 13 boards here | |||
IntelligenceLike-for-likePublic boards, shown as published, kept out of the index11
| Benchmark | GPT-5.2 | Qwen3.8 27B | Leader |
|---|---|---|---|
| AA Intelligence Index | 30.4 | 33.7 | Qwen3.8 27B |
| Epoch Capabilities Index | 153.7 | 151.7 | GPT-5.2 |
Agentic workDirectional onlyPublic boards, shown as published, kept out of the index22 boards, none shared
| Benchmark | GPT-5.2 | Qwen3.8 27B | Leader |
|---|---|---|---|
| AA Agentic Index | Not measured | 45.8 | Only Qwen3.8 27B measured |
| Design Arena: agents building web apps | 1086 | Not measured | Only GPT-5.2 measured |
| Design Arena: agents building full-stack apps | 1043 | Not measured | Only GPT-5.2 measured |
| Design Arena: agents building mobile apps | 1093 | Not measured | Only GPT-5.2 measured |
| Design Arena: agents building Android apps | 1059 | Not measured | Only GPT-5.2 measured |
| Design Arena: agents building Godot games | 1142 | Not measured | Only GPT-5.2 measured |
| LiveBench research | Not measured | 59.4% | Only Qwen3.8 27B measured |
| LiveBench browser | Not measured | 32.0% | Only Qwen3.8 27B measured |
| LiveBench long agentic | Not measured | 27.9% | Only Qwen3.8 27B measured |
| Finance Agent (Vals AI) | Not measured | 48.5% | Only Qwen3.8 27B measured |
| Tax Agent Bench (Vals AI) | Not measured | 30.9% | Only Qwen3.8 27B measured |
| Legal Research Bench (Vals AI) | Not measured | 36.1% | Only Qwen3.8 27B measured |
| Excel Modeling Benchmark (Vals AI) | Not measured | 59.7% | Only Qwen3.8 27B measured |
| Terminal-Bench-Science (Vals AI) | Not measured | 0.0% | Only Qwen3.8 27B measured |
| τ²-Bench Telecom | 84.8% | Not measured | Only GPT-5.2 measured |
| APEX-Agents | Not measured | 47.5% | Only Qwen3.8 27B measured |
| METR time horizons | 75.3% | Not measured | Only GPT-5.2 measured |
| Remote Labor Index | 2.5% | Not measured | Only GPT-5.2 measured |
| GDPval | 49.7% | Not measured | Only GPT-5.2 measured |
| DeepResearch Bench | 41.1% | Not measured | Only GPT-5.2 measured |
| Vending-Bench 2 | $3,591 | Not measured | Only GPT-5.2 measured |
| Berkeley Function Calling Leaderboard (BFCL) | 55.9% | Not measured | Only GPT-5.2 measured |
| Neither model has a result on the other 19 boards here | |||
ReasoningLike-for-likePublic boards, shown as published, kept out of the index47
| Benchmark | GPT-5.2 | Qwen3.8 27B | Leader |
|---|---|---|---|
| GPQA Diamond | 91.7% | 88.9% | GPT-5.2 |
| LiveBench reasoning | Not measured | 80.0% | Only Qwen3.8 27B measured |
| MMMU Pro (Vals AI) | 86.7% | 83.9% | GPT-5.2 |
| GPQA Diamond (Artificial Analysis run) | 90.3% | 90.5% | Qwen3.8 27B |
| Humanity's Last Exam (Artificial Analysis run) | 37.7% | 33.9% | GPT-5.2 |
| GPQA Diamond (Epoch AI run) | 91.4% | Not measured | Only GPT-5.2 measured |
| ARC-AGI | 86.2% | 87.5% | Qwen3.8 27B |
| ARC-AGI-2 | 52.9% | 42.4% | GPT-5.2 |
| SimpleBench | 45.8% | Not measured | Only GPT-5.2 measured |
| Chess puzzles | 48.0% | Not measured | Only GPT-5.2 measured |
| Mystery game puzzles | 23.0% | Not measured | Only GPT-5.2 measured |
| EBR-bench | 23.0% | Not measured | Only GPT-5.2 measured |
| VPCT | 84.0% | Not measured | Only GPT-5.2 measured |
| LMArena Text: hard prompts | 1428 | 1459 | Qwen3.8 27B |
| LMArena Text: expert | 1445 | 1479 | Qwen3.8 27B |
| LMArena Text: hard prompts in English | 1431 | 1461 | Qwen3.8 27B |
| LMArena Vision: diagrams | 1265 | 1285 | Qwen3.8 27B |
| LMArena Vision: homework | 1276 | 1295 | Qwen3.8 27B |
| Kagi LLM Benchmark | 73.3% | Not measured | Only GPT-5.2 measured |
| Neither model has a result on the other 3 boards here | |||
MathLike-for-likePublic boards, shown as published, kept out of the index03
| Benchmark | GPT-5.2 | Qwen3.8 27B | Leader |
|---|---|---|---|
| AIME | 96.9% | Not measured | Only GPT-5.2 measured |
| MGSM (Vals AI) | 94.0% | Not measured | Only GPT-5.2 measured |
| AA Math Index | 99.0 | Not measured | Only GPT-5.2 measured |
| AIME 2025 (Artificial Analysis run) | 99.0% | Not measured | Only GPT-5.2 measured |
| OTIS Mock AIME | 96.1% | Not measured | Only GPT-5.2 measured |
| FrontierMath, tiers 1 to 3 | 67.4% | Not measured | Only GPT-5.2 measured |
| FrontierMath, tier 4 | 31.7% | Not measured | Only GPT-5.2 measured |
| ProofBench | 15.0% | 16.0% | Qwen3.8 27B |
| LMArena Text: math | 1440 | 1455 | Qwen3.8 27B |
| LMArena Text: mathematical work | 1442 | 1458 | Qwen3.8 27B |
| Neither model has a result on the other 3 boards here | |||
KnowledgeLike-for-likePublic boards, shown as published, kept out of the index54
| Benchmark | GPT-5.2 | Qwen3.8 27B | Leader |
|---|---|---|---|
| MMLU Pro | 86.2% | 84.3% | GPT-5.2 |
| LegalBench (Vals AI) | 82.8% | 82.4% | GPT-5.2 |
| CaseLaw (Vals AI) | 66.0% | Not measured | Only GPT-5.2 measured |
| CorpFin (Vals AI) | 65.9% | Not measured | Only GPT-5.2 measured |
| TaxEval (Vals AI) | 75.8% | 70.8% | GPT-5.2 |
| MortgageTax (Vals AI) | 67.1% | 64.9% | GPT-5.2 |
| MedQA (Vals AI) | 94.1% | Not measured | Only GPT-5.2 measured |
| MedCode (Vals AI) | 49.8% | 28.7% | GPT-5.2 |
| MMLU Pro (Artificial Analysis run) | 87.4% | Not measured | Only GPT-5.2 measured |
| SimpleQA Verified | 37.1% | Not measured | Only GPT-5.2 measured |
| LMArena Text: science | 1425 | 1467 | Qwen3.8 27B |
| LMArena Text: medicine and healthcare | 1425 | 1437 | Qwen3.8 27B |
| LMArena Text: legal and government | 1427 | 1447 | Qwen3.8 27B |
| LMArena Text: business and finance | 1414 | 1438 | Qwen3.8 27B |
| FACTS Benchmark Suite | 48.5% | Not measured | Only GPT-5.2 measured |
| Neither model has a result on the other 3 boards here | |||
Web designDirectional onlyPublic boards, shown as published, kept out of the index9 boards, none shared
| Benchmark | GPT-5.2 | Qwen3.8 27B | Leader |
|---|---|---|---|
| Design Arena: websites | 1195 | Not measured | Only GPT-5.2 measured |
| Design Arena: UI components | 1190 | Not measured | Only GPT-5.2 measured |
| Design Arena: game dev | 1195 | Not measured | Only GPT-5.2 measured |
| Design Arena: data viz | 1202 | Not measured | Only GPT-5.2 measured |
| Design Arena: 3D | 1079 | Not measured | Only GPT-5.2 measured |
| Design Arena: all code categories | 1170 | Not measured | Only GPT-5.2 measured |
| Design Arena: SVG | 1139 | Not measured | Only GPT-5.2 measured |
| Design Arena: ASCII art | 1152 | Not measured | Only GPT-5.2 measured |
| LiveBench design | Not measured | 75.9% | Only Qwen3.8 27B measured |
GeneralLike-for-likePublic boards, shown as published, kept out of the index622
| Benchmark | GPT-5.2 | Qwen3.8 27B | Leader |
|---|---|---|---|
| LiveBench data analysis | Not measured | 76.6% | Only Qwen3.8 27B measured |
| SAGE (Vals AI) | 49.3% | 52.4% | Qwen3.8 27B |
| MedScribe (Vals AI) | 84.4% | 83.8% | GPT-5.2 |
| IFBench | 75.4% | Not measured | Only GPT-5.2 measured |
| Long context reasoning (AA-LCR) | 82.7% | 82.0% | GPT-5.2 |
| DTBench | 90.9% | 88.0% | GPT-5.2 |
| LMCA | 43.9% | 41.4% | GPT-5.2 |
| CL-bench | 18.2% | Not measured | Only GPT-5.2 measured |
| Surface Evolver Bench | Not measured | 45.0% | Only Qwen3.8 27B measured |
| Furniture assembly | 38.3% | Not measured | Only GPT-5.2 measured |
| LMArena Text | 1416 | 1440 | Qwen3.8 27B |
| LMArena Text: instruction following | 1409 | 1436 | Qwen3.8 27B |
| LMArena Text: creative writing | 1376 | 1383 | Qwen3.8 27B |
| LMArena Vision | 1246 | 1269 | Qwen3.8 27B |
| LMArena Text: English | 1422 | 1448 | Qwen3.8 27B |
| LMArena Text: non-English | 1405 | 1428 | Qwen3.8 27B |
| LMArena Text: Chinese | 1447 | 1501 | Qwen3.8 27B |
| LMArena Text: German | 1444 | 1443 | GPT-5.2 |
| LMArena Text: Japanese | 1420 | 1387 | GPT-5.2 |
| LMArena Text: Korean | 1374 | 1384 | Qwen3.8 27B |
| LMArena Text: French | 1454 | 1465 | Qwen3.8 27B |
| LMArena Text: Spanish | 1413 | 1449 | Qwen3.8 27B |
| LMArena Text: Russian | 1415 | 1415 | Qwen3.8 27B |
| LMArena Text: Polish | 1426 | 1439 | Qwen3.8 27B |
| LMArena Text: multi-turn | 1419 | 1440 | Qwen3.8 27B |
| LMArena Text: longer queries | 1413 | 1448 | Qwen3.8 27B |
| LMArena Text: writing and language | 1392 | 1405 | Qwen3.8 27B |
| LMArena Text: entertainment and media | 1378 | 1393 | Qwen3.8 27B |
| LMArena Vision: English | 1243 | 1267 | Qwen3.8 27B |
| LMArena Vision: Chinese | 1305 | 1326 | Qwen3.8 27B |
| LMArena Vision: reading text | 1257 | 1281 | Qwen3.8 27B |
| LMArena Vision: humor | 1227 | 1235 | Qwen3.8 27B |
| LMArena Vision: captioning | 1267 | Not measured | Only GPT-5.2 measured |
| LMArena Vision: recognizing things | 1232 | Not measured | Only GPT-5.2 measured |
| LMArena Vision: creative writing | 1225 | 1237 | Qwen3.8 27B |
| LMArena Document | 1405 | Not measured | Only GPT-5.2 measured |
| Neither model has a result on the other 10 boards here | |||
Image generationNot comparablePublic boards, shown as published, kept out of the index0 boards, none shared
| Benchmark | GPT-5.2 | Qwen3.8 27B | 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
Qwen3.8 27B costs less
Repository review
50K input + 3K output tokens
Qwen3.8 27B costs less
Agent loop
200K input + 10K output tokens
Qwen3.8 27B 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 | OpenAI GPT-5.2 | Qwen3.8 27B |
|---|---|---|
| Maker | OpenAI | Qwen |
| Weights | No public weights listed | Open weights |
| Released | December 10, 2025 | August 14, 2026 |
| Context window | 400K | 1M |
| Input $/Mtok | $1.75 | $0.425 |
| Output $/Mtok | $14.00 | $2.55 |
| Coding index | 55.9 | 74.2 |
| Independent measurements behind it | 3 | 4 |
| Cost per solved task | $1.91 | $1.26 |
| Source | OpenRouter listing | OpenRouter listing |
Questions
Is GPT-5.2 better than Qwen3.8 27B for coding?
Pick Qwen3.8 27B for coding: it scores 74.2 on the Vibe Coding Index against 55.9 for GPT-5.2. Pick GPT-5.2 for work measured by LiveCodeBench (Vals AI) or Epoch Capabilities Index, where it leads.
Which is cheaper, GPT-5.2 or Qwen3.8 27B?
GPT-5.2 costs $1.75 per million input tokens and $14.00 per million output tokens; Qwen3.8 27B costs $0.425 and $2.55. Per solved SWE-bench Verified task, GPT-5.2 costs $1.91 and Qwen3.8 27B $1.26.
Which has the bigger context window, GPT-5.2 or Qwen3.8 27B?
Qwen3.8 27B, with 1M tokens against 400K.
How many benchmarks is the GPT-5.2 vs Qwen3.8 27B comparison based on?
60 benchmarks have a result for both models as of October 22, 2026. GPT-5.2's coding index rests on 3 independent benchmarks and Qwen3.8 27B's on 4. Where only one model has a result, the page says so instead of naming a winner.