GPT-5.4 Mini vs Qwen3.8 Max (0902)
Updated October 21, 2026. 76 benchmarks have a result for both. Coding is ranked on the Vibe Coding Index; every other board is shown as published.
Decision reading
Pick Qwen3.8 Max (0902) for coding: it scores 63.5 on the Vibe Coding Index against 28.2 for GPT-5.4 Mini.
Pick GPT-5.4 Mini when output tokens cost 1.3x less, or when a solved task costs $0.70 against $1.32.
- Shared results
- 76
- GPT-5.4 Mini only
- 12
- Qwen3.8 Max (0902) only
- 37
- Like-for-like categories
- 7 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.
Coding
Vibe Coding Index, four public coding benchmarks
Qwen3.8 Max (0902)
Qwen3.8 Max (0902) scores 63.5 against 28.2.
1 and 2 benchmarks behind the scores
Intelligence
AA Intelligence Index
Qwen3.8 Max (0902)
Qwen3.8 Max (0902) leads, 45.4 against 24.1.
Same board, both measured
Agentic work
AA Agentic Index
Qwen3.8 Max (0902)
Qwen3.8 Max (0902) leads, 56.0 against 17.9.
Same board, both measured
Reasoning
GPQA Diamond
Qwen3.8 Max (0902)
Qwen3.8 Max (0902) leads, 93.7% against 83.1%.
Same board, both measured
Math
AIME
No call
Only GPT-5.4 Mini has a AIME result, 95.6%.
Directional only
Knowledge
MMLU Pro
Qwen3.8 Max (0902)
Qwen3.8 Max (0902) leads, 88.6% against 84.5%.
Same board, both measured
Web design
Design Arena: websites
No call
Only Qwen3.8 Max (0902) has a Design Arena: websites result, 1275.
Directional only
Long documents
Listed context window
Qwen3.8 Max (0902)
Qwen3.8 Max (0902) takes 1M tokens against 400K.
Listed by OpenRouter
Everyday cost
1K input + 500 output tokens
GPT-5.4 Mini
GPT-5.4 Mini $0.0030, Qwen3.8 Max (0902) $0.0050 per turn.
List prices
Cost per solved task
One SWE-bench Verified task, solved
GPT-5.4 Mini
GPT-5.4 Mini $0.70, Qwen3.8 Max (0902) $1.32.
Published attempt cost and pass rate
At a glance
- Coding index
- GPT-5.4 Mini 28.2 from 1 benchmark, Qwen3.8 Max (0902) 63.5 from 2 benchmarks.
- Widest category
- Agentic work, 23.3 points apart on average over 5 shared benchmarks.
- Widest single gap
- LMArena WebDev: GPT-5.4 Mini 1397 against Qwen3.8 Max (0902) 1674, 39.5 points apart on a 0 to 100 scale.
- Output price
- GPT-5.4 Mini $4.50 against Qwen3.8 Max (0902) $6.00 per million output tokens, 1.3x apart.
- Context window
- GPT-5.4 Mini takes 400K tokens, Qwen3.8 Max (0902) 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-5.4 Mini | Bars | Qwen3.8 Max (0902) | Ties |
|---|---|---|---|---|
| Coding | 0 | 13 | 1 tie | |
| Intelligence | 0 | 2 | ||
| Agentic work | 0 | 5 | ||
| Reasoning | 0 | 12 | ||
| Math | 0 | 6 | ||
| Knowledge | 0 | 9 | ||
| General | 0 | 28 | ||
| All shared boards | 0 | 75 | 1 tie |
CodingLike-for-likeFour of these boards feed the Vibe Coding Index013
| Benchmark | GPT-5.4 Mini | Qwen3.8 Max (0902) | Leader |
|---|---|---|---|
| LiveBench coding | Not measured | 60.3% | Only Qwen3.8 Max (0902) measured |
| LMArena WebDev | 1397 | 1674 | Qwen3.8 Max (0902) |
| AA Coding Index | 56.1 | 76.2 | Qwen3.8 Max (0902) |
| LiveCodeBench (Vals AI) | 81.5% | 87.8% | Qwen3.8 Max (0902) |
| SWE-bench Verified (Vals AI) | 73.0% | 85.6% | Qwen3.8 Max (0902) |
| Terminal-Bench 4 (Vals AI) | 2.5% | 34.3% | Qwen3.8 Max (0902) |
| Vibe Code Bench (Vals AI) | 48.0% | 64.7% | Qwen3.8 Max (0902) |
| Code migration (Vals AI) | 12.9% | 24.0% | Qwen3.8 Max (0902) |
| ProgramBench (Vals AI) | 0.0% | 0.0% | Tie |
| IOI (Vals AI) | Not measured | 68.9% | Only Qwen3.8 Max (0902) measured |
| Terminal-Bench 2.1 (Vals AI) | 54.7% | 67.4% | Qwen3.8 Max (0902) |
| SWE-bench Multimodal (Vals AI) | Not measured | 42.5% | Only Qwen3.8 Max (0902) measured |
| Vibe Code Bench 1-100 (Vals AI) | Not measured | 12.8% | Only Qwen3.8 Max (0902) measured |
| SkillsBench (Vals AI) | Not measured | 42.0% | Only Qwen3.8 Max (0902) measured |
| CyberBench (Vals AI) | Not measured | 28.6% | Only Qwen3.8 Max (0902) measured |
| SciCode | 52.1% | 53.2% | Qwen3.8 Max (0902) |
| Terminal-Bench Hard (Artificial Analysis run) | 52.3% | Not measured | Only GPT-5.4 Mini measured |
| FrontierCode | 27.0% | Not measured | Only GPT-5.4 Mini measured |
| FrontierSWE | Not measured | 17.8% | Only Qwen3.8 Max (0902) measured |
| WeirdML | 60.3% | Not measured | Only GPT-5.4 Mini measured |
| ALE-Bench | 1188.6 | Not measured | Only GPT-5.4 Mini measured |
| LMArena Text: coding | 1438 | 1502 | Qwen3.8 Max (0902) |
| LMArena Text: software and IT | 1436 | 1498 | Qwen3.8 Max (0902) |
| LMArena WebDev: React | 1386 | 1692 | Qwen3.8 Max (0902) |
| LMArena WebDev: HTML | 1419 | 1649 | Qwen3.8 Max (0902) |
| LMArena WebDev: from an image | Not measured | 1637 | Only Qwen3.8 Max (0902) measured |
| Neither model has a result on the other 12 boards here | |||
IntelligenceLike-for-likePublic boards, shown as published, kept out of the index02
| Benchmark | GPT-5.4 Mini | Qwen3.8 Max (0902) | Leader |
|---|---|---|---|
| AA Intelligence Index | 24.1 | 45.4 | Qwen3.8 Max (0902) |
| Epoch Capabilities Index | 148.7 | 157.4 | Qwen3.8 Max (0902) |
Agentic workLike-for-likePublic boards, shown as published, kept out of the index05
| Benchmark | GPT-5.4 Mini | Qwen3.8 Max (0902) | Leader |
|---|---|---|---|
| AA Agentic Index | 17.9 | 56.0 | Qwen3.8 Max (0902) |
| Design Arena: agents building web apps | Not measured | 1301 | Only Qwen3.8 Max (0902) measured |
| Design Arena: agents building full-stack apps | Not measured | 1305 | Only Qwen3.8 Max (0902) measured |
| Design Arena: agents building mobile apps | Not measured | 1223 | Only Qwen3.8 Max (0902) measured |
| Design Arena: agents building Android apps | Not measured | 1207 | Only Qwen3.8 Max (0902) measured |
| Design Arena: agents building Godot games | Not measured | 1234 | Only Qwen3.8 Max (0902) measured |
| Design Arena: PowerPoint slides | Not measured | 1257 | Only Qwen3.8 Max (0902) measured |
| Design Arena: spreadsheets | Not measured | 868 | Only Qwen3.8 Max (0902) measured |
| LiveBench research | Not measured | 69.9% | Only Qwen3.8 Max (0902) measured |
| LiveBench browser | Not measured | 32.0% | Only Qwen3.8 Max (0902) measured |
| LiveBench long agentic | Not measured | 42.3% | Only Qwen3.8 Max (0902) measured |
| Finance Agent (Vals AI) | 45.4% | 50.6% | Qwen3.8 Max (0902) |
| Tax Agent Bench (Vals AI) | 8.8% | 32.1% | Qwen3.8 Max (0902) |
| Legal Research Bench (Vals AI) | 12.5% | 47.6% | Qwen3.8 Max (0902) |
| Excel Modeling Benchmark (Vals AI) | 45.4% | 60.1% | Qwen3.8 Max (0902) |
| Terminal-Bench-Science (Vals AI) | Not measured | 12.9% | Only Qwen3.8 Max (0902) measured |
| τ²-Bench Telecom | 83.3% | Not measured | Only GPT-5.4 Mini measured |
| APEX-Agents | Not measured | 63.3% | Only Qwen3.8 Max (0902) measured |
| DeepResearch Bench | 36.3% | Not measured | Only GPT-5.4 Mini measured |
| LMArena Search | Not measured | 1261 | Only Qwen3.8 Max (0902) measured |
| Neither model has a result on the other 19 boards here | |||
ReasoningLike-for-likePublic boards, shown as published, kept out of the index012
| Benchmark | GPT-5.4 Mini | Qwen3.8 Max (0902) | Leader |
|---|---|---|---|
| GPQA Diamond | 83.1% | 93.7% | Qwen3.8 Max (0902) |
| LiveBench reasoning | Not measured | 88.2% | Only Qwen3.8 Max (0902) measured |
| MysteryMechanism (Vals AI) | Not measured | 23.9% | Only Qwen3.8 Max (0902) measured |
| MMMU Pro (Vals AI) | 79.3% | 88.0% | Qwen3.8 Max (0902) |
| GPQA Diamond (Artificial Analysis run) | 87.5% | 92.8% | Qwen3.8 Max (0902) |
| Humanity's Last Exam (Artificial Analysis run) | 28.1% | 43.1% | Qwen3.8 Max (0902) |
| GPQA Diamond (Epoch AI run) | 86.9% | 92.7% | Qwen3.8 Max (0902) |
| ARC-AGI | 63.7% | Not measured | Only GPT-5.4 Mini measured |
| ARC-AGI-2 | 18.9% | Not measured | Only GPT-5.4 Mini measured |
| Chess puzzles | 23.0% | 46.0% | Qwen3.8 Max (0902) |
| Mystery game puzzles | 11.0% | 38.0% | Qwen3.8 Max (0902) |
| LMArena Text: hard prompts | 1424 | 1496 | Qwen3.8 Max (0902) |
| LMArena Text: expert | 1435 | 1507 | Qwen3.8 Max (0902) |
| LMArena Text: hard prompts in English | 1420 | 1503 | Qwen3.8 Max (0902) |
| LMArena Vision: diagrams | 1266 | 1323 | Qwen3.8 Max (0902) |
| LMArena Vision: homework | 1285 | 1336 | Qwen3.8 Max (0902) |
| Neither model has a result on the other 5 boards here | |||
MathLike-for-likePublic boards, shown as published, kept out of the index06
| Benchmark | GPT-5.4 Mini | Qwen3.8 Max (0902) | Leader |
|---|---|---|---|
| AIME | 95.6% | Not measured | Only GPT-5.4 Mini measured |
| OTIS Mock AIME | 88.9% | 100.0% | Qwen3.8 Max (0902) |
| FrontierMath, tiers 1 to 3 | 51.2% | 74.7% | Qwen3.8 Max (0902) |
| FrontierMath, tier 4 | 9.8% | 46.3% | Qwen3.8 Max (0902) |
| ProofBench | 21.0% | 58.0% | Qwen3.8 Max (0902) |
| LMArena Text: math | 1418 | 1501 | Qwen3.8 Max (0902) |
| LMArena Text: mathematical work | 1429 | 1491 | Qwen3.8 Max (0902) |
| Neither model has a result on the other 6 boards here | |||
KnowledgeLike-for-likePublic boards, shown as published, kept out of the index09
| Benchmark | GPT-5.4 Mini | Qwen3.8 Max (0902) | Leader |
|---|---|---|---|
| MMLU Pro | 84.5% | 88.6% | Qwen3.8 Max (0902) |
| LegalBench (Vals AI) | Not measured | 83.6% | Only Qwen3.8 Max (0902) measured |
| CaseLaw (Vals AI) | 51.7% | Not measured | Only GPT-5.4 Mini measured |
| CorpFin (Vals AI) | 60.9% | 65.8% | Qwen3.8 Max (0902) |
| TaxEval (Vals AI) | 71.2% | 75.5% | Qwen3.8 Max (0902) |
| MortgageTax (Vals AI) | 63.5% | 64.0% | Qwen3.8 Max (0902) |
| MedCode (Vals AI) | Not measured | 40.7% | Only Qwen3.8 Max (0902) measured |
| SimpleQA Verified | 29.4% | 47.3% | Qwen3.8 Max (0902) |
| LMArena Text: science | 1417 | 1512 | Qwen3.8 Max (0902) |
| LMArena Text: medicine and healthcare | 1395 | 1503 | Qwen3.8 Max (0902) |
| LMArena Text: legal and government | 1414 | 1492 | Qwen3.8 Max (0902) |
| LMArena Text: business and finance | 1418 | 1480 | Qwen3.8 Max (0902) |
| Neither model has a result on the other 5 boards here | |||
Web designDirectional onlyPublic boards, shown as published, kept out of the index8 boards, none shared
| Benchmark | GPT-5.4 Mini | Qwen3.8 Max (0902) | Leader |
|---|---|---|---|
| Design Arena: websites | Not measured | 1275 | Only Qwen3.8 Max (0902) measured |
| Design Arena: UI components | Not measured | 1340 | Only Qwen3.8 Max (0902) measured |
| Design Arena: game dev | Not measured | 1313 | Only Qwen3.8 Max (0902) measured |
| Design Arena: data viz | Not measured | 1281 | Only Qwen3.8 Max (0902) measured |
| Design Arena: 3D | Not measured | 1337 | Only Qwen3.8 Max (0902) measured |
| Design Arena: all code categories | Not measured | 1293 | Only Qwen3.8 Max (0902) measured |
| Design Arena: ASCII art | Not measured | 1201 | Only Qwen3.8 Max (0902) measured |
| LiveBench design | Not measured | 78.6% | Only Qwen3.8 Max (0902) measured |
| Neither model has a result on the other 1 board here | |||
GeneralLike-for-likePublic boards, shown as published, kept out of the index028
| Benchmark | GPT-5.4 Mini | Qwen3.8 Max (0902) | Leader |
|---|---|---|---|
| LiveBench data analysis | Not measured | 78.4% | Only Qwen3.8 Max (0902) measured |
| Public Benefits Bench (Vals AI) | Not measured | 67.1% | Only Qwen3.8 Max (0902) measured |
| SAGE (Vals AI) | 50.8% | 51.2% | Qwen3.8 Max (0902) |
| MedScribe (Vals AI) | Not measured | 85.0% | Only Qwen3.8 Max (0902) measured |
| Vals Multimodal Index | 54.2% | 65.4% | Qwen3.8 Max (0902) |
| IFBench | 73.3% | Not measured | Only GPT-5.4 Mini measured |
| Long context reasoning (AA-LCR) | 77.0% | 80.3% | Qwen3.8 Max (0902) |
| DTBench | 80.0% | 92.0% | Qwen3.8 Max (0902) |
| LMCA | 40.8% | 46.2% | Qwen3.8 Max (0902) |
| Furniture assembly | Not measured | 20.0% | Only Qwen3.8 Max (0902) measured |
| LMArena Text | 1412 | 1483 | Qwen3.8 Max (0902) |
| LMArena Text: instruction following | 1405 | 1481 | Qwen3.8 Max (0902) |
| LMArena Text: creative writing | 1370 | 1477 | Qwen3.8 Max (0902) |
| LMArena Vision | 1245 | 1313 | Qwen3.8 Max (0902) |
| LMArena Text: English | 1412 | 1491 | Qwen3.8 Max (0902) |
| LMArena Text: non-English | 1405 | 1473 | Qwen3.8 Max (0902) |
| LMArena Text: Chinese | 1446 | 1539 | Qwen3.8 Max (0902) |
| LMArena Text: German | 1408 | 1478 | Qwen3.8 Max (0902) |
| LMArena Text: Japanese | 1375 | 1473 | Qwen3.8 Max (0902) |
| LMArena Text: Korean | 1370 | 1460 | Qwen3.8 Max (0902) |
| LMArena Text: French | 1441 | 1498 | Qwen3.8 Max (0902) |
| LMArena Text: Spanish | 1405 | 1489 | Qwen3.8 Max (0902) |
| LMArena Text: Russian | 1417 | 1482 | Qwen3.8 Max (0902) |
| LMArena Text: Polish | 1415 | 1471 | Qwen3.8 Max (0902) |
| LMArena Text: multi-turn | 1429 | 1489 | Qwen3.8 Max (0902) |
| LMArena Text: longer queries | 1408 | 1488 | Qwen3.8 Max (0902) |
| LMArena Text: writing and language | 1393 | 1477 | Qwen3.8 Max (0902) |
| LMArena Text: entertainment and media | 1373 | 1464 | Qwen3.8 Max (0902) |
| LMArena Vision: English | 1244 | 1311 | Qwen3.8 Max (0902) |
| LMArena Vision: Chinese | 1286 | 1358 | Qwen3.8 Max (0902) |
| LMArena Vision: reading text | 1253 | 1322 | Qwen3.8 Max (0902) |
| LMArena Vision: humor | 1235 | 1304 | Qwen3.8 Max (0902) |
| LMArena Vision: recognizing things | 1218 | Not measured | Only GPT-5.4 Mini measured |
| LMArena Vision: creative writing | 1222 | 1312 | Qwen3.8 Max (0902) |
| 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
| Benchmark | GPT-5.4 Mini | Qwen3.8 Max (0902) | 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
GPT-5.4 Mini costs less
Repository review
50K input + 3K output tokens
GPT-5.4 Mini costs less
Agent loop
200K input + 10K output tokens
GPT-5.4 Mini 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-5.4 Mini | Qwen3.8 Max (0902) |
|---|---|---|
| Maker | OpenAI | Qwen |
| Weights | No public weights listed | No public weights listed |
| Released | March 17, 2026 | September 3, 2026 |
| Context window | 400K | 1M |
| Input $/Mtok | $0.75 | $2.00 |
| Output $/Mtok | $4.50 | $6.00 |
| Coding index | 28.2 | 63.5 |
| Independent measurements behind it | 1 | 2 |
| Cost per solved task | $0.70 | $1.32 |
| Source | OpenRouter listing | OpenRouter listing |
Questions
Is GPT-5.4 Mini better than Qwen3.8 Max (0902) for coding?
Pick Qwen3.8 Max (0902) for coding: it scores 63.5 on the Vibe Coding Index against 28.2 for GPT-5.4 Mini. Pick GPT-5.4 Mini when output tokens cost 1.3x less, or when a solved task costs $0.70 against $1.32.
Which is cheaper, GPT-5.4 Mini or Qwen3.8 Max (0902)?
GPT-5.4 Mini costs $0.75 per million input tokens and $4.50 per million output tokens; Qwen3.8 Max (0902) costs $2.00 and $6.00. Per solved SWE-bench Verified task, GPT-5.4 Mini costs $0.70 and Qwen3.8 Max (0902) $1.32.
Which has the bigger context window, GPT-5.4 Mini or Qwen3.8 Max (0902)?
Qwen3.8 Max (0902), with 1M tokens against 400K.
How many benchmarks is the GPT-5.4 Mini vs Qwen3.8 Max (0902) comparison based on?
76 benchmarks have a result for both models as of October 21, 2026. GPT-5.4 Mini's coding index rests on 1 independent benchmark and Qwen3.8 Max (0902)'s on 2. Where only one model has a result, the page says so instead of naming a winner.