Enter opens the highlighted result. Esc closes.

GPT-5 Nano vs Qwen3 30B A3B Instruct 2507

GPT-5 Nano leads Qwen3 30B A3B Instruct 2507 on the Artificial Analysis Intelligence Index, 13.0 to 7.5, and Qwen3 30B A3B Instruct 2507 costs 1.3x less per output token ($0.30 vs $0.40 per million). Updated October 22, 2026; 43 benchmarks have a result for both.

Replace GPT-5 Nano Replace Qwen3 30B A3B Instruct 2507
GPT-5 NanoOpenAI
Coding
No data
Overall
35.6
Output $/M
$0.40
10
Qwen3 30B A3B Instruct 2507Qwen
Coding
No data
Overall
41.5
Output $/M
$0.30
33
vs boards won of 43 shared

GPT-5 Nano or Qwen3 30B A3B Instruct 2507: which to pick

Neither model has a coding score yet, so the two cannot be ranked on coding.

Pick GPT-5 Nano for a larger context window (400K against 262K tokens) and work measured by LMArena Text: coding or LMArena Text: software and IT, where it leads.

Model AGPT-5 NanoOpenAINot benchmarkedNo coding benchmark$0.109 per solved taskModel BQwen3 30B A3B Instruct 2507QwenNot benchmarkedNo coding benchmarkNo cost per solved task
Shared results
43
GPT-5 Nano only
45
Qwen3 30B A3B Instruct 2507 only
2
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-5 Nano not benchmarked, Qwen3 30B A3B Instruct 2507 not benchmarked.
Widest category
Math, 16.1 points apart on average over 5 shared benchmarks.
Widest single gap
OTIS Mock AIME: GPT-5 Nano 81.1% against Qwen3 30B A3B Instruct 2507 62.2%, 18.9 points apart on a 0 to 100 scale.
Output price
GPT-5 Nano $0.40 against Qwen3 30B A3B Instruct 2507 $0.30 per million output tokens, 1.3x apart.
Context window
GPT-5 Nano takes 400K tokens, Qwen3 30B A3B Instruct 2507 262K.

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-5 NanoBarsQwen3 30B A3B Instruct 2507Ties
Coding22
Intelligence10
Agentic work20
Reasoning15
Math32
Knowledge05
General119
All shared boards1033
CodingLike-for-likeSix of these boards feed the Vibe Coding Index22
Coding benchmarks for GPT-5 Nano and Qwen3 30B A3B Instruct 2507
BenchmarkGPT-5 NanoQwen3 30B A3B Instruct 2507Leader
SWE-bench Verified34.8%Not measuredOnly GPT-5 Nano measured
LiveCodeBench (Vals AI)70.2%Not measuredOnly GPT-5 Nano measured
LiveCodeBench (Artificial Analysis run)78.9%70.7%GPT-5 Nano
SciCodeNot measured33.0%Only Qwen3 30B A3B Instruct 2507 measured
Terminal-Bench Hard (Artificial Analysis run)17.4%6.1%GPT-5 Nano
Terminal-Bench21.8%Not measuredOnly GPT-5 Nano measured
WeirdML38.1%Not measuredOnly GPT-5 Nano measured
LMArena Text: coding13511416Qwen3 30B A3B Instruct 2507
LMArena Text: software and IT13451417Qwen3 30B A3B Instruct 2507
Neither model has a result on the other 34 boards here
IntelligenceLike-for-likePublic boards, shown as published, kept out of the index10
Intelligence benchmarks for GPT-5 Nano and Qwen3 30B A3B Instruct 2507
BenchmarkGPT-5 NanoQwen3 30B A3B Instruct 2507Leader
AA Intelligence Index13.07.5GPT-5 Nano
Epoch Capabilities Index139.2Not measuredOnly GPT-5 Nano measured
Agentic workLike-for-likePublic boards, shown as published, kept out of the index20
Agentic work benchmarks for GPT-5 Nano and Qwen3 30B A3B Instruct 2507
BenchmarkGPT-5 NanoQwen3 30B A3B Instruct 2507Leader
τ²-Bench Telecom36.5%28.1%GPT-5 Nano
Berkeley Function Calling Leaderboard (BFCL)51.4%41.4%GPT-5 Nano
Neither model has a result on the other 40 boards here
ReasoningLike-for-likePublic boards, shown as published, kept out of the index15
Reasoning benchmarks for GPT-5 Nano and Qwen3 30B A3B Instruct 2507
BenchmarkGPT-5 NanoQwen3 30B A3B Instruct 2507Leader
GPQA Diamond63.4%Not measuredOnly GPT-5 Nano measured
MMMU Pro (Vals AI)70.9%Not measuredOnly GPT-5 Nano measured
GPQA Diamond (Artificial Analysis run)67.6%70.7%Qwen3 30B A3B Instruct 2507
Humanity's Last Exam (Artificial Analysis run)9.5%10.3%Qwen3 30B A3B Instruct 2507
GPQA Diamond (Epoch AI run)69.4%55.6%GPT-5 Nano
ARC-AGI20.7%Not measuredOnly GPT-5 Nano measured
ARC-AGI-22.6%Not measuredOnly GPT-5 Nano measured
Chess puzzles22.0%Not measuredOnly GPT-5 Nano measured
Mystery game puzzles9.0%Not measuredOnly GPT-5 Nano measured
VPCT37.2%Not measuredOnly GPT-5 Nano measured
LMArena Text: hard prompts13281398Qwen3 30B A3B Instruct 2507
LMArena Text: expert13211397Qwen3 30B A3B Instruct 2507
LMArena Text: hard prompts in English13381406Qwen3 30B A3B Instruct 2507
LMArena Vision: diagrams1141Not measuredOnly GPT-5 Nano measured
LMArena Vision: homework1157Not measuredOnly GPT-5 Nano measured
Kagi LLM Benchmark62.2%Not measuredOnly GPT-5 Nano measured
Neither model has a result on the other 7 boards here
MathLike-for-likePublic boards, shown as published, kept out of the index32
Math benchmarks for GPT-5 Nano and Qwen3 30B A3B Instruct 2507
BenchmarkGPT-5 NanoQwen3 30B A3B Instruct 2507Leader
AIME81.2%Not measuredOnly GPT-5 Nano measured
MATH 500 (Vals AI)93.8%Not measuredOnly GPT-5 Nano measured
MGSM (Vals AI)89.3%Not measuredOnly GPT-5 Nano measured
AA Math Index83.766.3GPT-5 Nano
AIME 2025 (Artificial Analysis run)83.7%66.3%GPT-5 Nano
OTIS Mock AIME81.1%62.2%GPT-5 Nano
FrontierMath, tiers 1 to 320.0%Not measuredOnly GPT-5 Nano measured
FrontierMath, tier 42.4%Not measuredOnly GPT-5 Nano measured
ProofBench12.0%Not measuredOnly GPT-5 Nano measured
MATH level 595.2%Not measuredOnly GPT-5 Nano measured
LMArena Text: math13171394Qwen3 30B A3B Instruct 2507
LMArena Text: mathematical work13261411Qwen3 30B A3B Instruct 2507
MathArena12.8%Not measuredOnly GPT-5 Nano measured
Neither model has a result on the other 1 board here
KnowledgeLike-for-likePublic boards, shown as published, kept out of the index05
Knowledge benchmarks for GPT-5 Nano and Qwen3 30B A3B Instruct 2507
BenchmarkGPT-5 NanoQwen3 30B A3B Instruct 2507Leader
MMLU Pro76.1%Not measuredOnly GPT-5 Nano measured
LegalBench (Vals AI)50.1%Not measuredOnly GPT-5 Nano measured
CaseLaw (Vals AI)52.6%Not measuredOnly GPT-5 Nano measured
TaxEval (Vals AI)67.4%Not measuredOnly GPT-5 Nano measured
MortgageTax (Vals AI)53.6%Not measuredOnly GPT-5 Nano measured
MedQA (Vals AI)93.3%Not measuredOnly GPT-5 Nano measured
MedCode (Vals AI)30.4%Not measuredOnly GPT-5 Nano measured
MMLU Pro (Artificial Analysis run)78.0%80.5%Qwen3 30B A3B Instruct 2507
SimpleQA Verified11.7%Not measuredOnly GPT-5 Nano measured
LMArena Text: science13181393Qwen3 30B A3B Instruct 2507
LMArena Text: medicine and healthcare13351392Qwen3 30B A3B Instruct 2507
LMArena Text: legal and government12981387Qwen3 30B A3B Instruct 2507
LMArena Text: business and finance13111389Qwen3 30B A3B Instruct 2507
Neither model has a result on the other 7 boards here
Web designDirectional onlyPublic boards, shown as published, kept out of the index6 boards, none shared
Web design benchmarks for GPT-5 Nano and Qwen3 30B A3B Instruct 2507
BenchmarkGPT-5 NanoQwen3 30B A3B Instruct 2507Leader
Design Arena: websites1101Not measuredOnly GPT-5 Nano measured
Design Arena: UI components1064Not measuredOnly GPT-5 Nano measured
Design Arena: game dev1044Not measuredOnly GPT-5 Nano measured
Design Arena: data viz1056Not measuredOnly GPT-5 Nano measured
Design Arena: 3D965Not measuredOnly GPT-5 Nano measured
Design Arena: all code categories1084Not measuredOnly GPT-5 Nano measured
Neither model has a result on the other 3 boards here
GeneralLike-for-likePublic boards, shown as published, kept out of the index119
General benchmarks for GPT-5 Nano and Qwen3 30B A3B Instruct 2507
BenchmarkGPT-5 NanoQwen3 30B A3B Instruct 2507Leader
SAGE (Vals AI)30.4%Not measuredOnly GPT-5 Nano measured
MedScribe (Vals AI)72.9%Not measuredOnly GPT-5 Nano measured
IFBench67.6%50.7%GPT-5 Nano
Long context reasoning (AA-LCR)45.0%61.3%Qwen3 30B A3B Instruct 2507
DTBench62.7%67.2%Qwen3 30B A3B Instruct 2507
LMCA7.9%22.4%Qwen3 30B A3B Instruct 2507
Fiction.LiveBench44.4%Not measuredOnly GPT-5 Nano measured
LMArena Text13201384Qwen3 30B A3B Instruct 2507
LMArena Text: instruction following13061363Qwen3 30B A3B Instruct 2507
LMArena Text: creative writing12491317Qwen3 30B A3B Instruct 2507
LMArena Vision1159Not measuredOnly GPT-5 Nano measured
LMArena Text: English13211391Qwen3 30B A3B Instruct 2507
LMArena Text: non-English13131372Qwen3 30B A3B Instruct 2507
LMArena Text: Chinese13561432Qwen3 30B A3B Instruct 2507
LMArena Text: German13271379Qwen3 30B A3B Instruct 2507
LMArena Text: Japanese12271337Qwen3 30B A3B Instruct 2507
LMArena Text: Korean12691330Qwen3 30B A3B Instruct 2507
LMArena Text: FrenchNot measured1419Only Qwen3 30B A3B Instruct 2507 measured
LMArena Text: Spanish13601404Qwen3 30B A3B Instruct 2507
LMArena Text: Russian12961369Qwen3 30B A3B Instruct 2507
LMArena Text: Polish13141353Qwen3 30B A3B Instruct 2507
LMArena Text: multi-turn13111378Qwen3 30B A3B Instruct 2507
LMArena Text: longer queries13121378Qwen3 30B A3B Instruct 2507
LMArena Text: writing and language12831336Qwen3 30B A3B Instruct 2507
LMArena Text: entertainment and media12681334Qwen3 30B A3B Instruct 2507
LMArena Vision: English1170Not measuredOnly GPT-5 Nano measured
LMArena Vision: Chinese1176Not measuredOnly GPT-5 Nano measured
LMArena Vision: reading text1152Not measuredOnly GPT-5 Nano measured
LMArena Vision: humor1110Not measuredOnly GPT-5 Nano 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
Image generation benchmarks for GPT-5 Nano and Qwen3 30B A3B Instruct 2507
BenchmarkGPT-5 NanoQwen3 30B A3B Instruct 2507Leader
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-5 Nano$0.0003Fits in one request
Qwen3 30B A3B Instruct 2507$0.0002Fits in one request

Qwen3 30B A3B Instruct 2507 costs less

Repository review

50K input + 3K output tokens

GPT-5 Nano$0.0037Fits in one request
Qwen3 30B A3B Instruct 2507$0.0054Fits in one request

GPT-5 Nano costs less

Agent loop

200K input + 10K output tokens

GPT-5 Nano$0.014Fits in one request
Qwen3 30B A3B Instruct 2507$0.021Fits in one request

GPT-5 Nano 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-5 Nano and Qwen3 30B A3B Instruct 2507
SpecificationOpenAI GPT-5 NanoQwen3 30B A3B Instruct 2507
MakerOpenAIQwen
WeightsNo public weights listedOpen weights
ReleasedAugust 7, 2025July 29, 2025
Context window400K262K
Input $/Mtok$0.05$0.09
Output $/Mtok$0.40$0.30
Coding indexNot benchmarkedNot benchmarked
Independent measurements behind it00
Cost per solved task$0.109Not computable
SourceOpenRouter listingOpenRouter listing

Questions

Is GPT-5 Nano better than Qwen3 30B A3B Instruct 2507 for coding?

Neither model has a coding score yet, so the two cannot be ranked on coding. Pick GPT-5 Nano for a larger context window (400K against 262K tokens) and work measured by LMArena Text: coding or LMArena Text: software and IT, where it leads.

Which is cheaper, GPT-5 Nano or Qwen3 30B A3B Instruct 2507?

GPT-5 Nano costs $0.05 per million input tokens and $0.40 per million output tokens; Qwen3 30B A3B Instruct 2507 costs $0.09 and $0.30. Cost per solved task is known for GPT-5 Nano only.

Which has the bigger context window, GPT-5 Nano or Qwen3 30B A3B Instruct 2507?

GPT-5 Nano, with 400K tokens against 262K.

How many benchmarks is the GPT-5 Nano vs Qwen3 30B A3B Instruct 2507 comparison based on?

43 benchmarks have a result for both models as of October 22, 2026. GPT-5 Nano's coding index rests on 0 independent benchmarks and Qwen3 30B A3B Instruct 2507's on 0. Where only one model has a result, the page says so instead of naming a winner.

Compare next