Enter opens the highlighted result. Esc closes.

GPT-6.2 Astra vs Qwen3.8 27B

GPT-6.2 Astra leads Qwen3.8 27B on coding, 87.0 to 69.7 on the Vibe Coding Index, and Qwen3.8 27B costs 15.7x less per output token ($2.55 vs $40.00 per million). Updated November 22, 2026; 43 benchmarks have a result for both.

GPT-6.2 AstraOpenAI
Coding
87.0
Overall
86.4
Output $/M
$40.00
29
Qwen3.8 27BQwen
Coding
69.7
Overall
51.7
Output $/M
$2.55
14
vs boards won of 43 shared

GPT-6.2 Astra or Qwen3.8 27B: which to pick

Pick GPT-6.2 Astra for coding: it scores 87.0 on the Vibe Coding Index against 69.7 for Qwen3.8 27B.

Pick Qwen3.8 27B for output tokens 15.7x cheaper ($2.55 vs $40.00 per million) and work measured by LMArena Text: coding or LMArena Text: software and IT, where it leads.

Shared results
43
GPT-6.2 Astra only
33
Qwen3.8 27B only
48
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.

At a glance

Coding index
GPT-6.2 Astra 87.0 from 4 benchmarks, Qwen3.8 27B 69.7 from 4 benchmarks.
Widest category
Agentic work, 38.0 points apart on average over 7 shared benchmarks.
Widest single gap
Terminal-Bench-Science (Vals AI): GPT-6.2 Astra 71.4% against Qwen3.8 27B 0.0%, 71.4 points apart on a 0 to 100 scale.
Output price
GPT-6.2 Astra $40.00 against Qwen3.8 27B $2.55 per million output tokens, 15.7x apart.
Context window
GPT-6.2 Astra 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.

Boards won, where both have a result
CategoryGPT-6.2 AstraBarsQwen3.8 27BTies
Coding82
Intelligence20
Agentic work61
Reasoning62
Knowledge13
Web design10
General56
All shared boards2914
CodingLike-for-likeSix of these boards feed the Vibe Coding Index82
Coding benchmarks for GPT-6.2 Astra and Qwen3.8 27B
BenchmarkGPT-6.2 AstraQwen3.8 27BLeader
LiveBench coding91.9%69.3%GPT-6.2 Astra
LMArena WebDev17661590GPT-6.2 Astra
LiveCodeBench (Vals AI)Not measured84.0%Only Qwen3.8 27B measured
SWE-bench Verified (Vals AI)Not measured86.0%Only Qwen3.8 27B measured
Terminal-Bench 4 (Vals AI)63.1%Not measuredOnly GPT-6.2 Astra measured
Vibe Code Bench (Vals AI)89.9%64.8%GPT-6.2 Astra
Code migration (Vals AI)Not measured14.2%Only Qwen3.8 27B measured
ProgramBench (Vals AI)Not measured0.0%Only Qwen3.8 27B measured
IOI (Vals AI)Not measured39.1%Only Qwen3.8 27B measured
Terminal-Bench 2.1 (Vals AI)Not measured58.4%Only Qwen3.8 27B measured
SkillsBench (Vals AI)Not measured38.1%Only Qwen3.8 27B measured
CyberBench (Vals AI)Not measured67.8%Only Qwen3.8 27B measured
SciCodeNot measured46.6%Only Qwen3.8 27B measured
Terminal-Bench 4.0 (Artificial Analysis run)64.1%5.6%GPT-6.2 Astra
FrontierCode 2.049.0%14.9%GPT-6.2 Astra
MirrorCode67.8%Not measuredOnly GPT-6.2 Astra measured
LMArena Text: coding14501481Qwen3.8 27B
LMArena Text: software and IT14431475Qwen3.8 27B
LMArena WebDev: React17701594GPT-6.2 Astra
LMArena WebDev: HTML18541571GPT-6.2 Astra
LMArena WebDev: from an imageNot measured1575Only Qwen3.8 27B measured
GameBench 297.243.4GPT-6.2 Astra
Neither model has a result on the other 30 boards here
IntelligenceLike-for-likePublic boards, shown as published, kept out of the index20
Intelligence benchmarks for GPT-6.2 Astra and Qwen3.8 27B
BenchmarkGPT-6.2 AstraQwen3.8 27BLeader
AA Intelligence Index62.120.1GPT-6.2 Astra
Epoch Capabilities Index172.8152.6GPT-6.2 Astra
Agentic workLike-for-likePublic boards, shown as published, kept out of the index61
Agentic work benchmarks for GPT-6.2 Astra and Qwen3.8 27B
BenchmarkGPT-6.2 AstraQwen3.8 27BLeader
Design Arena: agents building web apps1334Not measuredOnly GPT-6.2 Astra measured
Design Arena: agents building full-stack apps1351Not measuredOnly GPT-6.2 Astra measured
Design Arena: agents building mobile apps1305Not measuredOnly GPT-6.2 Astra measured
Design Arena: agents building Android apps1257Not measuredOnly GPT-6.2 Astra measured
Design Arena: spreadsheets1278Not measuredOnly GPT-6.2 Astra measured
LiveBench research91.5%61.3%GPT-6.2 Astra
LiveBench browser93.0%24.2%GPT-6.2 Astra
LiveBench long agentic70.8%15.9%GPT-6.2 Astra
Finance Agent (Vals AI)Not measured48.5%Only Qwen3.8 27B measured
Tax Agent Bench (Vals AI)21.8%30.9%Qwen3.8 27B
Legal Research Bench (Vals AI)42.8%36.1%GPT-6.2 Astra
Harvey's Legal Agent Benchmark (Vals AI)16.3%Not measuredOnly GPT-6.2 Astra measured
Excel Modeling Benchmark (Vals AI)Not measured59.7%Only Qwen3.8 27B measured
BioMysteryBench (Vals AI)80.7%Not measuredOnly GPT-6.2 Astra measured
Terminal-Bench-Science (Vals AI)71.4%0.0%GPT-6.2 Astra
Vals Science Index68.4%Not measuredOnly GPT-6.2 Astra measured
Robo-bench (Vals AI)32.7%Not measuredOnly GPT-6.2 Astra measured
APEX-Agents72.2%47.5%GPT-6.2 Astra
SpireClimb41.3%Not measuredOnly GPT-6.2 Astra measured
LMArena Search1191Not measuredOnly GPT-6.2 Astra measured
Neither model has a result on the other 24 boards here
ReasoningLike-for-likePublic boards, shown as published, kept out of the index62
Reasoning benchmarks for GPT-6.2 Astra and Qwen3.8 27B
BenchmarkGPT-6.2 AstraQwen3.8 27BLeader
GPQA DiamondNot measured88.9%Only Qwen3.8 27B measured
LiveBench reasoning91.6%80.0%GPT-6.2 Astra
MysteryMechanism (Vals AI)53.1%Not measuredOnly GPT-6.2 Astra measured
MMMU Pro (Vals AI)Not measured83.9%Only Qwen3.8 27B measured
GPQA Diamond (Artificial Analysis run)Not measured90.5%Only Qwen3.8 27B measured
Humanity's Last Exam (Artificial Analysis run)Not measured33.9%Only Qwen3.8 27B measured
HLE-Diamond (Artificial Analysis run)72.2%15.3%GPT-6.2 Astra
CritPt v2, hard subset (Artificial Analysis run)71.5%16.5%GPT-6.2 Astra
GPQA Diamond (Epoch AI run)96.8%Not measuredOnly GPT-6.2 Astra measured
ARC-AGI98.5%87.5%GPT-6.2 Astra
ARC-AGI-295.8%42.4%GPT-6.2 Astra
SimpleBench77.3%60.2%GPT-6.2 Astra
Chess puzzles78.0%Not measuredOnly GPT-6.2 Astra measured
Mystery game puzzles95.0%Not measuredOnly GPT-6.2 Astra measured
EBR-bench92.4%Not measuredOnly GPT-6.2 Astra measured
LMArena Text: hard prompts14251458Qwen3.8 27B
LMArena Text: expertNot measured1479Only Qwen3.8 27B measured
LMArena Text: hard prompts in English14261463Qwen3.8 27B
LMArena Vision: diagramsNot measured1280Only Qwen3.8 27B measured
LMArena Vision: homeworkNot measured1288Only Qwen3.8 27B measured
Humanity's Last Exam (Scale run)66.9%Not measuredOnly GPT-6.2 Astra measured
GoBench2493Not measuredOnly GPT-6.2 Astra measured
Neither model has a result on the other 4 boards here
MathDirectional onlyPublic boards, shown as published, kept out of the index7 boards, none shared
Math benchmarks for GPT-6.2 Astra and Qwen3.8 27B
BenchmarkGPT-6.2 AstraQwen3.8 27BLeader
OTIS Mock AIME100.0%Not measuredOnly GPT-6.2 Astra measured
FrontierMath, tiers 1 to 392.6%Not measuredOnly GPT-6.2 Astra measured
FrontierMath, tier 4100.0%Not measuredOnly GPT-6.2 Astra measured
ProofBenchNot measured16.0%Only Qwen3.8 27B measured
LMArena Text: mathNot measured1459Only Qwen3.8 27B measured
LMArena Text: mathematical workNot measured1462Only Qwen3.8 27B measured
MathArena92.9%Not measuredOnly GPT-6.2 Astra measured
Neither model has a result on the other 7 boards here
KnowledgeLike-for-likePublic boards, shown as published, kept out of the index13
Knowledge benchmarks for GPT-6.2 Astra and Qwen3.8 27B
BenchmarkGPT-6.2 AstraQwen3.8 27BLeader
MMLU ProNot measured84.3%Only Qwen3.8 27B measured
LegalBench (Vals AI)Not measured82.4%Only Qwen3.8 27B measured
TaxEval (Vals AI)Not measured70.8%Only Qwen3.8 27B measured
MortgageTax (Vals AI)Not measured64.9%Only Qwen3.8 27B measured
MedCode (Vals AI)Not measured28.7%Only Qwen3.8 27B measured
AA-Omniscience50.0-6.4GPT-6.2 Astra
SimpleQA Verified78.5%Not measuredOnly GPT-6.2 Astra measured
GeoBenchNot measured38.0%Only Qwen3.8 27B measured
LMArena Text: science14151465Qwen3.8 27B
LMArena Text: medicine and healthcareNot measured1439Only Qwen3.8 27B measured
LMArena Text: legal and government14151447Qwen3.8 27B
LMArena Text: business and finance14131440Qwen3.8 27B
Neither model has a result on the other 9 boards here
Web designLike-for-likePublic boards, shown as published, kept out of the index10
Web design benchmarks for GPT-6.2 Astra and Qwen3.8 27B
BenchmarkGPT-6.2 AstraQwen3.8 27BLeader
Design Arena: websites1333Not measuredOnly GPT-6.2 Astra measured
Design Arena: UI components1341Not measuredOnly GPT-6.2 Astra measured
Design Arena: game dev1501Not measuredOnly GPT-6.2 Astra measured
Design Arena: 3D1430Not measuredOnly GPT-6.2 Astra measured
Design Arena: all code categories1389Not measuredOnly GPT-6.2 Astra measured
LiveBench design82.8%72.4%GPT-6.2 Astra
Neither model has a result on the other 3 boards here
GeneralLike-for-likePublic boards, shown as published, kept out of the index56
General benchmarks for GPT-6.2 Astra and Qwen3.8 27B
BenchmarkGPT-6.2 AstraQwen3.8 27BLeader
Vals Index67.7%Not measuredOnly GPT-6.2 Astra measured
LiveBench data analysis84.1%76.6%GPT-6.2 Astra
MedScribe (Vals AI)Not measured83.8%Only Qwen3.8 27B measured
Long context reasoning (AA-LCR)Not measured82.0%Only Qwen3.8 27B measured
GDPval-AANot measured1426Only Qwen3.8 27B measured
DTBenchNot measured88.0%Only Qwen3.8 27B measured
LMCANot measured41.4%Only Qwen3.8 27B measured
Surface Evolver BenchNot measured45.0%Only Qwen3.8 27B measured
Furniture assembly89.2%Not measuredOnly GPT-6.2 Astra measured
LMArena Text14141437Qwen3.8 27B
LMArena Text: instruction following14231433Qwen3.8 27B
LMArena Text: creative writing14171378GPT-6.2 Astra
LMArena VisionNot measured1271Only Qwen3.8 27B measured
LMArena Text: English14171448Qwen3.8 27B
LMArena Text: non-English14041425Qwen3.8 27B
LMArena Text: ChineseNot measured1491Only Qwen3.8 27B measured
LMArena Text: GermanNot measured1437Only Qwen3.8 27B measured
LMArena Text: JapaneseNot measured1390Only Qwen3.8 27B measured
LMArena Text: KoreanNot measured1398Only Qwen3.8 27B measured
LMArena Text: FrenchNot measured1466Only Qwen3.8 27B measured
LMArena Text: SpanishNot measured1445Only Qwen3.8 27B measured
LMArena Text: Russian14371412GPT-6.2 Astra
LMArena Text: PolishNot measured1449Only Qwen3.8 27B measured
LMArena Text: multi-turn14251439Qwen3.8 27B
LMArena Text: longer queries14311446Qwen3.8 27B
LMArena Text: writing and language14401402GPT-6.2 Astra
LMArena Text: entertainment and media14131389GPT-6.2 Astra
LMArena Vision: EnglishNot measured1275Only Qwen3.8 27B measured
LMArena Vision: ChineseNot measured1327Only Qwen3.8 27B measured
LMArena Vision: reading textNot measured1279Only Qwen3.8 27B measured
LMArena Vision: humorNot measured1242Only Qwen3.8 27B measured
LMArena Vision: creative writingNot measured1238Only Qwen3.8 27B measured
MultiChallenge73.2%Not measuredOnly GPT-6.2 Astra measured
Neither model has a result on the other 16 boards here
Image generationNot comparablePublic boards, shown as published, kept out of the index0 boards, none shared
Image generation benchmarks for GPT-6.2 Astra and Qwen3.8 27B
BenchmarkGPT-6.2 AstraQwen3.8 27BLeader
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

GPT-6.2 Astra$0.028Fits in one request
Qwen3.8 27B$0.0017Fits in one request

Qwen3.8 27B costs less

Repository review

50K input + 3K output tokens

GPT-6.2 Astra$0.520Fits in one request
Qwen3.8 27B$0.029Fits in one request

Qwen3.8 27B costs less

Agent loop

200K input + 10K output tokens

GPT-6.2 Astra$2.00Fits in one request
Qwen3.8 27B$0.111Fits in one request

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.

Specifications of GPT-6.2 Astra and Qwen3.8 27B
SpecificationOpenAI GPT-6.2 AstraQwen3.8 27B
MakerOpenAIQwen
WeightsNo public weights listedOpen weights
ReleasedNovember 17, 2026August 14, 2026
Context window1.05M1.02M
Input $/Mtok$8.00$0.425
Output $/Mtok$40.00$2.55
Coding index87.069.7
Independent measurements behind it44
Cost per solved taskNot computable$1.26
SourceOpenRouter listingOpenRouter listing

Questions

Is GPT-6.2 Astra better than Qwen3.8 27B for coding?

Pick GPT-6.2 Astra for coding: it scores 87.0 on the Vibe Coding Index against 69.7 for Qwen3.8 27B. Pick Qwen3.8 27B for output tokens 15.7x cheaper ($2.55 vs $40.00 per million) and work measured by LMArena Text: coding or LMArena Text: software and IT, where it leads.

Which is cheaper, GPT-6.2 Astra or Qwen3.8 27B?

GPT-6.2 Astra costs $8.00 per million input tokens and $40.00 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 Astra or Qwen3.8 27B?

GPT-6.2 Astra, with 1.05M tokens against 1.02M.

How many benchmarks is the GPT-6.2 Astra vs Qwen3.8 27B comparison based on?

43 benchmarks have a result for both models as of November 22, 2026. GPT-6.2 Astra'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.

Compare next

Embed this comparison on your site

Free under CC BY 4.0, while the source link stays visible. Preview the frame · Terms and other embeds

<iframe src="https://llmbenchmark.ai/embed/compare/gpt-6-2-astra-vs-qwen3-8-27b" title="GPT-6.2 Astra vs Qwen3.8 27B, LLMBenchmark.ai" width="100%" height="360" style="border:0;max-width:100%" loading="lazy"></iframe>
<p style="font:13px/1.4 sans-serif;margin:6px 0 0">Source: <a href="https://llmbenchmark.ai/compare/gpt-6-2-astra-vs-qwen3-8-27b">LLMBenchmark.ai, GPT-6.2 Astra vs Qwen3.8 27B</a></p>