GPT-6.2 Sol vs Qwen3.8 27B
GPT-6.2 Sol leads Qwen3.8 27B on coding, 83.1 to 69.7 on the Vibe Coding Index, and Qwen3.8 27B costs 3.9x less per output token ($2.55 vs $10.00 per million). Updated November 22, 2026; 42 benchmarks have a result for both.
GPT-6.2 Sol or Qwen3.8 27B: which to pick
Pick GPT-6.2 Sol for coding: it scores 83.1 on the Vibe Coding Index against 69.7 for Qwen3.8 27B.
Pick Qwen3.8 27B for output tokens 3.9x cheaper ($2.55 vs $10.00 per million) and work measured by LMArena Text: coding or LMArena Text: software and IT, where it leads.
- Shared results
- 42
- GPT-6.2 Sol only
- 14
- Qwen3.8 27B only
- 49
- 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
GPT-6.2 Sol
GPT-6.2 Sol scores 83.1 against 69.7.
4 benchmarks behind each score
Intelligence
AA Intelligence Index
GPT-6.2 Sol
GPT-6.2 Sol leads, 57.3 against 20.1.
Same board, both measured
Reasoning
GPQA Diamond
No call
Only Qwen3.8 27B has a GPQA Diamond result, 88.9%.
Directional only
Knowledge
MMLU Pro
No call
Only Qwen3.8 27B has a MMLU Pro result, 84.3%.
Directional only
General
Vals Index
No call
Only GPT-6.2 Sol has a Vals Index result, 66.1%.
Directional only
Long documents
Listed context window
GPT-6.2 Sol
GPT-6.2 Sol takes 1.05M tokens against 1.02M.
Listed by OpenRouter
Everyday cost
1K input + 500 output tokens
Qwen3.8 27B
GPT-6.2 Sol $0.0070, Qwen3.8 27B $0.0017 per turn.
List prices
Cost per solved task
One SWE-bench Verified task, solved
No call
Only Qwen3.8 27B has the number, $1.26.
Directional only
At a glance
- Coding index
- GPT-6.2 Sol 83.1 from 4 benchmarks, Qwen3.8 27B 69.7 from 4 benchmarks.
- Widest category
- Intelligence, 37.2 points apart on average over 1 shared benchmark.
- Widest single gap
- LiveBench browser: GPT-6.2 Sol 88.2% against Qwen3.8 27B 24.2%, 64.0 points apart on a 0 to 100 scale.
- Output price
- GPT-6.2 Sol $10.00 against Qwen3.8 27B $2.55 per million output tokens, 3.9x apart.
- Context window
- GPT-6.2 Sol takes 1.05M tokens, Qwen3.8 27B 1.02M.
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-6.2 Sol | Bars | Qwen3.8 27B | Ties |
|---|---|---|---|---|
| Coding | 8 | 2 | ||
| Intelligence | 1 | 0 | ||
| Agentic work | 6 | 1 | ||
| Reasoning | 6 | 2 | ||
| Knowledge | 1 | 2 | ||
| Web design | 1 | 0 | ||
| General | 5 | 7 | ||
| All shared boards | 28 | 14 |
CodingLike-for-likeSix of these boards feed the Vibe Coding Index82
| Benchmark | GPT-6.2 Sol | Qwen3.8 27B | Leader |
|---|---|---|---|
| LiveBench coding | 83.7% | 69.3% | GPT-6.2 Sol |
| LMArena WebDev | 1754 | 1590 | GPT-6.2 Sol |
| LiveCodeBench (Vals AI) | Not measured | 84.0% | Only Qwen3.8 27B measured |
| SWE-bench Verified (Vals AI) | Not measured | 86.0% | Only Qwen3.8 27B measured |
| Terminal-Bench 4 (Vals AI) | 60.6% | Not measured | Only GPT-6.2 Sol measured |
| Vibe Code Bench (Vals AI) | 87.6% | 64.8% | GPT-6.2 Sol |
| 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 |
| SciCode | Not measured | 46.6% | Only Qwen3.8 27B measured |
| Terminal-Bench 4.0 (Artificial Analysis run) | 60.1% | 5.6% | GPT-6.2 Sol |
| FrontierCode 2.0 | 44.4% | 14.9% | GPT-6.2 Sol |
| LMArena Text: coding | 1440 | 1481 | Qwen3.8 27B |
| LMArena Text: software and IT | 1434 | 1475 | Qwen3.8 27B |
| LMArena WebDev: React | 1761 | 1594 | GPT-6.2 Sol |
| LMArena WebDev: HTML | 1840 | 1571 | GPT-6.2 Sol |
| LMArena WebDev: from an image | Not measured | 1575 | Only Qwen3.8 27B measured |
| GameBench 2 | 94.4 | 43.4 | GPT-6.2 Sol |
| Neither model has a result on the other 31 boards here | |||
IntelligenceLike-for-likePublic boards, shown as published, kept out of the index10
| Benchmark | GPT-6.2 Sol | Qwen3.8 27B | Leader |
|---|---|---|---|
| AA Intelligence Index | 57.3 | 20.1 | GPT-6.2 Sol |
| Epoch Capabilities Index | Not measured | 152.6 | Only Qwen3.8 27B measured |
Agentic workLike-for-likePublic boards, shown as published, kept out of the index61
| Benchmark | GPT-6.2 Sol | Qwen3.8 27B | Leader |
|---|---|---|---|
| LiveBench research | 90.1% | 61.3% | GPT-6.2 Sol |
| LiveBench browser | 88.2% | 24.2% | GPT-6.2 Sol |
| LiveBench long agentic | 63.6% | 15.9% | GPT-6.2 Sol |
| Finance Agent (Vals AI) | Not measured | 48.5% | Only Qwen3.8 27B measured |
| Tax Agent Bench (Vals AI) | 24.8% | 30.9% | Qwen3.8 27B |
| Legal Research Bench (Vals AI) | 37.0% | 36.1% | GPT-6.2 Sol |
| Harvey's Legal Agent Benchmark (Vals AI) | 12.1% | Not measured | Only GPT-6.2 Sol measured |
| Excel Modeling Benchmark (Vals AI) | Not measured | 59.7% | Only Qwen3.8 27B measured |
| BioMysteryBench (Vals AI) | 77.8% | Not measured | Only GPT-6.2 Sol measured |
| Terminal-Bench-Science (Vals AI) | 58.6% | 0.0% | GPT-6.2 Sol |
| Vals Science Index | 57.8% | Not measured | Only GPT-6.2 Sol measured |
| APEX-Agents | 66.6% | 47.5% | GPT-6.2 Sol |
| LMArena Search | 1179 | Not measured | Only GPT-6.2 Sol measured |
| Neither model has a result on the other 31 boards here | |||
ReasoningLike-for-likePublic boards, shown as published, kept out of the index62
| Benchmark | GPT-6.2 Sol | Qwen3.8 27B | Leader |
|---|---|---|---|
| GPQA Diamond | Not measured | 88.9% | Only Qwen3.8 27B measured |
| LiveBench reasoning | 90.6% | 80.0% | GPT-6.2 Sol |
| MysteryMechanism (Vals AI) | 36.9% | Not measured | Only GPT-6.2 Sol measured |
| MMMU Pro (Vals AI) | Not measured | 83.9% | Only Qwen3.8 27B measured |
| GPQA Diamond (Artificial Analysis run) | Not measured | 90.5% | Only Qwen3.8 27B measured |
| Humanity's Last Exam (Artificial Analysis run) | Not measured | 33.9% | Only Qwen3.8 27B measured |
| HLE-Diamond (Artificial Analysis run) | 64.0% | 15.3% | GPT-6.2 Sol |
| CritPt v2, hard subset (Artificial Analysis run) | 64.0% | 16.5% | GPT-6.2 Sol |
| ARC-AGI | 97.5% | 87.5% | GPT-6.2 Sol |
| ARC-AGI-2 | 85.8% | 42.4% | GPT-6.2 Sol |
| SimpleBench | 69.9% | 60.2% | GPT-6.2 Sol |
| LMArena Text: hard prompts | 1419 | 1458 | Qwen3.8 27B |
| LMArena Text: expert | Not measured | 1479 | Only Qwen3.8 27B measured |
| LMArena Text: hard prompts in English | 1413 | 1463 | Qwen3.8 27B |
| LMArena Vision: diagrams | Not measured | 1280 | Only Qwen3.8 27B measured |
| LMArena Vision: homework | Not measured | 1288 | Only Qwen3.8 27B measured |
| Kagi LLM Benchmark | 67.0% | Not measured | Only GPT-6.2 Sol measured |
| Humanity's Last Exam (Scale run) | 56.5% | Not measured | Only GPT-6.2 Sol measured |
| Neither model has a result on the other 8 boards here | |||
MathDirectional onlyPublic boards, shown as published, kept out of the index4 boards, none shared
| Benchmark | GPT-6.2 Sol | Qwen3.8 27B | Leader |
|---|---|---|---|
| ProofBench | Not measured | 16.0% | Only Qwen3.8 27B measured |
| LMArena Text: math | Not measured | 1459 | Only Qwen3.8 27B measured |
| LMArena Text: mathematical work | Not measured | 1462 | Only Qwen3.8 27B measured |
| MathArena | 91.2% | Not measured | Only GPT-6.2 Sol measured |
| Neither model has a result on the other 10 boards here | |||
KnowledgeLike-for-likePublic boards, shown as published, kept out of the index12
| Benchmark | GPT-6.2 Sol | Qwen3.8 27B | Leader |
|---|---|---|---|
| MMLU Pro | Not measured | 84.3% | Only Qwen3.8 27B measured |
| LegalBench (Vals AI) | Not measured | 82.4% | Only Qwen3.8 27B measured |
| TaxEval (Vals AI) | Not measured | 70.8% | Only Qwen3.8 27B measured |
| MortgageTax (Vals AI) | Not measured | 64.9% | Only Qwen3.8 27B measured |
| MedCode (Vals AI) | Not measured | 28.7% | Only Qwen3.8 27B measured |
| AA-Omniscience | 40.4 | -6.4 | GPT-6.2 Sol |
| GeoBench | Not measured | 38.0% | Only Qwen3.8 27B measured |
| LMArena Text: science | 1432 | 1465 | Qwen3.8 27B |
| LMArena Text: medicine and healthcare | Not measured | 1439 | Only Qwen3.8 27B measured |
| LMArena Text: legal and government | Not measured | 1447 | Only Qwen3.8 27B measured |
| LMArena Text: business and finance | 1383 | 1440 | Qwen3.8 27B |
| Neither model has a result on the other 10 boards here | |||
Web designLike-for-likePublic boards, shown as published, kept out of the index10
| Benchmark | GPT-6.2 Sol | Qwen3.8 27B | Leader |
|---|---|---|---|
| Design Arena: game dev | 1437 | Not measured | Only GPT-6.2 Sol measured |
| Design Arena: 3D | 1482 | Not measured | Only GPT-6.2 Sol measured |
| Design Arena: all code categories | 1405 | Not measured | Only GPT-6.2 Sol measured |
| LiveBench design | 80.0% | 72.4% | GPT-6.2 Sol |
| Neither model has a result on the other 5 boards here | |||
GeneralLike-for-likePublic boards, shown as published, kept out of the index57
| Benchmark | GPT-6.2 Sol | Qwen3.8 27B | Leader |
|---|---|---|---|
| Vals Index | 66.1% | Not measured | Only GPT-6.2 Sol measured |
| LiveBench data analysis | 81.1% | 76.6% | GPT-6.2 Sol |
| MedScribe (Vals AI) | Not measured | 83.8% | Only Qwen3.8 27B measured |
| Long context reasoning (AA-LCR) | Not measured | 82.0% | Only Qwen3.8 27B measured |
| GDPval-AA | Not measured | 1426 | Only Qwen3.8 27B measured |
| DTBench | Not measured | 88.0% | Only Qwen3.8 27B measured |
| LMCA | Not measured | 41.4% | Only Qwen3.8 27B measured |
| Surface Evolver Bench | Not measured | 45.0% | Only Qwen3.8 27B measured |
| LMArena Text | 1414 | 1437 | Qwen3.8 27B |
| LMArena Text: instruction following | 1434 | 1433 | GPT-6.2 Sol |
| LMArena Text: creative writing | 1403 | 1378 | GPT-6.2 Sol |
| LMArena Vision | Not measured | 1271 | Only Qwen3.8 27B measured |
| LMArena Text: English | 1419 | 1448 | Qwen3.8 27B |
| LMArena Text: non-English | 1405 | 1425 | Qwen3.8 27B |
| LMArena Text: Chinese | 1434 | 1491 | Qwen3.8 27B |
| LMArena Text: German | Not measured | 1437 | Only Qwen3.8 27B measured |
| LMArena Text: Japanese | Not measured | 1390 | Only Qwen3.8 27B measured |
| LMArena Text: Korean | Not measured | 1398 | Only Qwen3.8 27B measured |
| LMArena Text: French | Not measured | 1466 | Only Qwen3.8 27B measured |
| LMArena Text: Spanish | Not measured | 1445 | Only Qwen3.8 27B measured |
| LMArena Text: Russian | 1401 | 1412 | Qwen3.8 27B |
| LMArena Text: Polish | Not measured | 1449 | Only Qwen3.8 27B measured |
| LMArena Text: multi-turn | 1420 | 1439 | Qwen3.8 27B |
| LMArena Text: longer queries | 1427 | 1446 | Qwen3.8 27B |
| LMArena Text: writing and language | 1420 | 1402 | GPT-6.2 Sol |
| LMArena Text: entertainment and media | 1408 | 1389 | GPT-6.2 Sol |
| LMArena Vision: English | Not measured | 1275 | Only Qwen3.8 27B measured |
| LMArena Vision: Chinese | Not measured | 1327 | Only Qwen3.8 27B measured |
| LMArena Vision: reading text | Not measured | 1279 | Only Qwen3.8 27B measured |
| LMArena Vision: humor | Not measured | 1242 | Only Qwen3.8 27B measured |
| LMArena Vision: creative writing | Not measured | 1238 | Only Qwen3.8 27B measured |
| MultiChallenge | 72.0% | Not measured | Only GPT-6.2 Sol measured |
| Neither model has a result on the other 17 boards here | |||
Image generationNot comparablePublic boards, shown as published, kept out of the index0 boards, none shared
| Benchmark | GPT-6.2 Sol | 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-6.2 Sol | Qwen3.8 27B |
|---|---|---|
| Maker | OpenAI | Qwen |
| Weights | No public weights listed | Open weights |
| Released | November 17, 2026 | August 14, 2026 |
| Context window | 1.05M | 1.02M |
| Input $/Mtok | $2.00 | $0.425 |
| Output $/Mtok | $10.00 | $2.55 |
| Coding index | 83.1 | 69.7 |
| Independent measurements behind it | 4 | 4 |
| Cost per solved task | Not computable | $1.26 |
| Source | OpenRouter listing | OpenRouter listing |
Questions
Is GPT-6.2 Sol better than Qwen3.8 27B for coding?
Pick GPT-6.2 Sol for coding: it scores 83.1 on the Vibe Coding Index against 69.7 for Qwen3.8 27B. Pick Qwen3.8 27B for output tokens 3.9x cheaper ($2.55 vs $10.00 per million) and work measured by LMArena Text: coding or LMArena Text: software and IT, where it leads.
Which is cheaper, GPT-6.2 Sol or Qwen3.8 27B?
GPT-6.2 Sol costs $2.00 per million input tokens and $10.00 per million output tokens; Qwen3.8 27B costs $0.425 and $2.55. Cost per solved task is known for Qwen3.8 27B only.
Which has the bigger context window, GPT-6.2 Sol or Qwen3.8 27B?
GPT-6.2 Sol, with 1.05M tokens against 1.02M.
How many benchmarks is the GPT-6.2 Sol vs Qwen3.8 27B comparison based on?
42 benchmarks have a result for both models as of November 22, 2026. GPT-6.2 Sol's coding index rests on 4 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.