LLMBenchmark.ai

Enter opens the highlighted result. Esc closes.

o3 vs Qwen3.5 397B A17B

Updated October 21, 2026. 58 benchmarks have a result for both. Coding is ranked on the Vibe Coding Index; every other board is shown as published.

Replace o3 Replace Qwen3.5 397B A17B
o3OpenAI
Coding
70.7
Overall
20.2
Output $/M
$8.00
5
Qwen3.5 397B A17BQwen
Coding
28.6
Overall
21.4
Output $/M
$2.34
53
vs boards won of 58 shared

Decision reading

Pick o3 for coding: it scores 70.7 on the Vibe Coding Index against 28.6 for Qwen3.5 397B A17B.

Pick Qwen3.5 397B A17B when output tokens cost 3.4x less, or when the context window is larger, 262K against 200K tokens.

Model Ao3OpenAI70.7/1002 independent benchmarks$0.571 per solved taskModel BQwen3.5 397B A17BQwen28.6/1001 independent benchmarkNo cost per solved task
Shared results
58
o3 only
39
Qwen3.5 397B A17B only
7
Like-for-like categories
8 of 9

Shared results can support a head-to-head reading. Results only one model has describe coverage, not superiority.

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
o3 70.7 from 2 benchmarks, Qwen3.5 397B A17B 28.6 from 1 benchmark.
Widest category
Agentic work, 14.9 points apart on average over 1 shared benchmark.
Widest single gap
Design Arena: all code categories: o3 1021 against Qwen3.5 397B A17B 1180, 19.9 points apart on a 0 to 100 scale.
Output price
o3 $8.00 against Qwen3.5 397B A17B $2.34 per million output tokens, 3.4x apart.
Context window
o3 takes 200K tokens, Qwen3.5 397B A17B 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
Categoryo3BarsQwen3.5 397B A17BTies
Coding13
Intelligence02
Agentic work01
Reasoning18
Math13
Knowledge04
Web design14
General128
All shared boards553
CodingLike-for-likeFour of these boards feed the Vibe Coding Index13
Coding benchmarks for o3 and Qwen3.5 397B A17B
Benchmarko3Qwen3.5 397B A17BLeader
Aider polyglot81.3%Not measuredOnly o3 measured
SWE-bench Verified58.4%Not measuredOnly o3 measured
SWE-bench Multimodal36.0%Not measuredOnly o3 measured
LMArena WebDevNot measured1400Only Qwen3.5 397B A17B measured
LiveCodeBench (Vals AI)83.9%Not measuredOnly o3 measured
LiveCodeBench (Artificial Analysis run)80.8%Not measuredOnly o3 measured
SciCodeNot measured44.8%Only Qwen3.5 397B A17B measured
Terminal-Bench Hard (Artificial Analysis run)37.1%40.9%Qwen3.5 397B A17B
GSO-Bench8.8%Not measuredOnly o3 measured
WeirdML52.4%Not measuredOnly o3 measured
ALE-Bench933.5621.9o3
LMArena Text: coding14071464Qwen3.5 397B A17B
LMArena Text: software and IT14151463Qwen3.5 397B A17B
LMArena WebDev: ReactNot measured1388Only Qwen3.5 397B A17B measured
LMArena WebDev: HTMLNot measured1410Only Qwen3.5 397B A17B measured
Neither model has a result on the other 23 boards here
IntelligenceLike-for-likePublic boards, shown as published, kept out of the index02
Intelligence benchmarks for o3 and Qwen3.5 397B A17B
Benchmarko3Qwen3.5 397B A17BLeader
AA Intelligence Index20.221.4Qwen3.5 397B A17B
Epoch Capabilities Index146.7146.8Qwen3.5 397B A17B
Agentic workLike-for-likePublic boards, shown as published, kept out of the index01
Agentic work benchmarks for o3 and Qwen3.5 397B A17B
Benchmarko3Qwen3.5 397B A17BLeader
τ²-Bench Telecom80.7%95.6%Qwen3.5 397B A17B
APEX-AgentsNot measured24.9%Only Qwen3.5 397B A17B measured
METR time horizons65.4%Not measuredOnly o3 measured
GDPval30.8%Not measuredOnly o3 measured
DeepResearch Bench45.2%Not measuredOnly o3 measured
OSWorld23.0%Not measuredOnly o3 measured
Neither model has a result on the other 33 boards here
ReasoningLike-for-likePublic boards, shown as published, kept out of the index18
Reasoning benchmarks for o3 and Qwen3.5 397B A17B
Benchmarko3Qwen3.5 397B A17BLeader
GPQA Diamond84.1%Not measuredOnly o3 measured
MMMU Pro (Vals AI)80.4%Not measuredOnly o3 measured
GPQA Diamond (Artificial Analysis run)82.7%89.3%Qwen3.5 397B A17B
Humanity's Last Exam (Artificial Analysis run)20.1%29.0%Qwen3.5 397B A17B
GPQA Diamond (Epoch AI run)81.8%86.4%Qwen3.5 397B A17B
ARC-AGI60.8%Not measuredOnly o3 measured
ARC-AGI-26.5%Not measuredOnly o3 measured
SimpleBench53.1%Not measuredOnly o3 measured
Chess puzzles37.0%Not measuredOnly o3 measured
Mystery game puzzles29.0%18.0%o3
VPCT52.0%Not measuredOnly o3 measured
LMArena Text: hard prompts14021448Qwen3.5 397B A17B
LMArena Text: expert14021462Qwen3.5 397B A17B
LMArena Text: hard prompts in English14111450Qwen3.5 397B A17B
LMArena Vision: diagrams12201277Qwen3.5 397B A17B
LMArena Vision: homework12341292Qwen3.5 397B A17B
Neither model has a result on the other 5 boards here
MathLike-for-likePublic boards, shown as published, kept out of the index13
Math benchmarks for o3 and Qwen3.5 397B A17B
Benchmarko3Qwen3.5 397B A17BLeader
AIME85.3%Not measuredOnly o3 measured
MATH 500 (Vals AI)94.6%Not measuredOnly o3 measured
MGSM (Vals AI)91.8%Not measuredOnly o3 measured
AA Math Index88.3Not measuredOnly o3 measured
AIME 2025 (Artificial Analysis run)88.3%Not measuredOnly o3 measured
OTIS Mock AIME84.4%88.9%Qwen3.5 397B A17B
FrontierMath, tiers 1 to 333.3%31.2%o3
MATH level 597.8%Not measuredOnly o3 measured
LMArena Text: math14261453Qwen3.5 397B A17B
LMArena Text: mathematical work14231454Qwen3.5 397B A17B
Neither model has a result on the other 3 boards here
KnowledgeLike-for-likePublic boards, shown as published, kept out of the index04
Knowledge benchmarks for o3 and Qwen3.5 397B A17B
Benchmarko3Qwen3.5 397B A17BLeader
MMLU Pro85.6%Not measuredOnly o3 measured
LegalBench (Vals AI)83.8%Not measuredOnly o3 measured
CorpFin (Vals AI)59.7%Not measuredOnly o3 measured
TaxEval (Vals AI)74.6%Not measuredOnly o3 measured
MortgageTax (Vals AI)65.7%Not measuredOnly o3 measured
MedQA (Vals AI)96.1%Not measuredOnly o3 measured
MedCode (Vals AI)47.3%Not measuredOnly o3 measured
MMLU Pro (Artificial Analysis run)85.3%Not measuredOnly o3 measured
SimpleQA Verified49.4%Not measuredOnly o3 measured
GeoBench74.0%Not measuredOnly o3 measured
LMArena Text: science14291457Qwen3.5 397B A17B
LMArena Text: medicine and healthcare14481455Qwen3.5 397B A17B
LMArena Text: legal and government14281441Qwen3.5 397B A17B
LMArena Text: business and finance13931435Qwen3.5 397B A17B
Neither model has a result on the other 3 boards here
Web designLike-for-likePublic boards, shown as published, kept out of the index14
Web design benchmarks for o3 and Qwen3.5 397B A17B
Benchmarko3Qwen3.5 397B A17BLeader
Design Arena: websites10361191Qwen3.5 397B A17B
Design Arena: UI components10161163Qwen3.5 397B A17B
Design Arena: game dev10341141Qwen3.5 397B A17B
Design Arena: data viz12001175o3
Design Arena: 3DNot measured1161Only Qwen3.5 397B A17B measured
Design Arena: all code categories10211180Qwen3.5 397B A17B
Design Arena: SVGNot measured1139Only Qwen3.5 397B A17B measured
Neither model has a result on the other 2 boards here
GeneralLike-for-likePublic boards, shown as published, kept out of the index128
General benchmarks for o3 and Qwen3.5 397B A17B
Benchmarko3Qwen3.5 397B A17BLeader
SAGE (Vals AI)41.8%Not measuredOnly o3 measured
MedScribe (Vals AI)76.7%Not measuredOnly o3 measured
IFBench71.4%78.8%Qwen3.5 397B A17B
Long context reasoning (AA-LCR)74.7%77.3%Qwen3.5 397B A17B
DTBench84.8%87.5%Qwen3.5 397B A17B
LMCA39.7%37.9%o3
Fiction.LiveBench88.9%Not measuredOnly o3 measured
Creative writing (Lech Mazur)83.9%Not measuredOnly o3 measured
CL-bench17.8%Not measuredOnly o3 measured
LMArena Text14101437Qwen3.5 397B A17B
LMArena Text: instruction following13681424Qwen3.5 397B A17B
LMArena Text: creative writing13591402Qwen3.5 397B A17B
LMArena Vision12141262Qwen3.5 397B A17B
LMArena Text: English14171443Qwen3.5 397B A17B
LMArena Text: non-English14011430Qwen3.5 397B A17B
LMArena Text: Chinese14371498Qwen3.5 397B A17B
LMArena Text: German14201445Qwen3.5 397B A17B
LMArena Text: Japanese14031429Qwen3.5 397B A17B
LMArena Text: Korean13701384Qwen3.5 397B A17B
LMArena Text: French14301462Qwen3.5 397B A17B
LMArena Text: Spanish13951441Qwen3.5 397B A17B
LMArena Text: Russian14061430Qwen3.5 397B A17B
LMArena Text: Polish14381447Qwen3.5 397B A17B
LMArena Text: multi-turn14051446Qwen3.5 397B A17B
LMArena Text: longer queries13721441Qwen3.5 397B A17B
LMArena Text: writing and language13721415Qwen3.5 397B A17B
LMArena Text: entertainment and media13731399Qwen3.5 397B A17B
LMArena Vision: English12211268Qwen3.5 397B A17B
LMArena Vision: Chinese12371314Qwen3.5 397B A17B
LMArena Vision: reading text12181271Qwen3.5 397B A17B
LMArena Vision: humor12221262Qwen3.5 397B A17B
LMArena Vision: captioning12101230Qwen3.5 397B A17B
LMArena Vision: recognizing things12341265Qwen3.5 397B A17B
LMArena Vision: creative writing12011260Qwen3.5 397B A17B
Neither model has a result on the other 12 boards here
Image generationNot comparablePublic boards, shown as published, kept out of the index0 boards, none shared
Image generation benchmarks for o3 and Qwen3.5 397B A17B
Benchmarko3Qwen3.5 397B A17BLeader
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

o3$0.0060Fits in one request
Qwen3.5 397B A17B$0.0016Fits in one request

Qwen3.5 397B A17B costs less

Repository review

50K input + 3K output tokens

o3$0.124Fits in one request
Qwen3.5 397B A17B$0.027Fits in one request

Qwen3.5 397B A17B costs less

Agent loop

200K input + 10K output tokens

o3$0.480Exceeds the context window
Qwen3.5 397B A17B$0.101Fits in one request

Qwen3.5 397B A17B 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 o3 and Qwen3.5 397B A17B
SpecificationOpenAI o3Qwen3.5 397B A17B
MakerOpenAIQwen
WeightsNo public weights listedOpen weights
ReleasedApril 16, 2025February 16, 2026
Context window200K262K
Input $/Mtok$2.00$0.39
Output $/Mtok$8.00$2.34
Coding index70.728.6
Independent measurements behind it21
Cost per solved task$0.571Not computable
SourceOpenRouter listingOpenRouter listing

Questions

Is o3 better than Qwen3.5 397B A17B for coding?

Pick o3 for coding: it scores 70.7 on the Vibe Coding Index against 28.6 for Qwen3.5 397B A17B. Pick Qwen3.5 397B A17B when output tokens cost 3.4x less, or when the context window is larger, 262K against 200K tokens.

Which is cheaper, o3 or Qwen3.5 397B A17B?

o3 costs $2.00 per million input tokens and $8.00 per million output tokens; Qwen3.5 397B A17B costs $0.39 and $2.34. Cost per solved task is known for o3 only.

Which has the bigger context window, o3 or Qwen3.5 397B A17B?

Qwen3.5 397B A17B, with 262K tokens against 200K.

How many benchmarks is the o3 vs Qwen3.5 397B A17B comparison based on?

58 benchmarks have a result for both models as of October 21, 2026. o3's coding index rests on 2 independent benchmarks and Qwen3.5 397B A17B's on 1. Where only one model has a result, the page says so instead of naming a winner.

Compare next