Model comparison
GPT-5.4 nano vs Qwen3.6 Max Preview
Qwen3.6 Max Preview is the stronger model overall, scoring 51.5 to 41.9 on the Noometry Index. GPT-5.4 nano costs 6.3× less per token, which makes it the better buy when Qwen3.6 Max Preview's lead doesn't matter for your workload.
Last verified . 24 shared benchmarks.
Alibaba (Qwen)Summary
- They share 24 benchmarks with published results for both. GPT-5.4 nano scores higher in 0 categories and Qwen3.6 Max Preview in 8 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where Qwen3.6 Max Preview leads 41.7 to 23.7.
- The biggest single-benchmark swing is SimpleQA Verified: 11.7% for GPT-5.4 nano and 52% for Qwen3.6 Max Preview.
- GPT-5.4 nano is cheaper at $0.20 / $1.25 per million input/output tokens, against $1.30 / $7.80 for Qwen3.6 Max Preview.
- GPT-5.4 nano accepts more context: 400K tokens versus 262K.
Side by side
| GPT-5.4 nano | Qwen3.6 Max Preview | |
|---|---|---|
| Provider | OpenAI | Alibaba (Qwen) |
| Noometry Index | 41.9 | 51.5 |
| Released | 2026-03-17 | 2026-04-20 |
| Weights | Proprietary | Proprietary |
| Context window | 400K | 262K |
| Max output | 128K | 66K |
| Input $ / M tokens | $0.20 | $1.30 |
| Output $ / M tokens | $1.25 | $7.80 |
| Results tracked | 40 | 29 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Qwen3.6 Max Preview leads
GPT-5.4 nano: 43.6 (#84), Qwen3.6 Max Preview: 48.7 (#54)
| Benchmark | GPT-5.4 nano | Qwen3.6 Max Preview |
|---|---|---|
| LMArena Coding | 1405 | 1471 |
| SWE-bench Verified | — | 76.7% |
| LMArena WebDev | — | 1482 |
| SciCode | 46.9% | — |
| WeirdML | 49.2% | — |
| ALE-Bench | 1,005 | — |
Agentic & Tool Use Not comparable
GPT-5.4 nano: —, Qwen3.6 Max Preview: —
| Benchmark | GPT-5.4 nano | Qwen3.6 Max Preview |
|---|---|---|
| Vending-Bench 2 | — | 4,254 |
Reasoning Qwen3.6 Max Preview leads
GPT-5.4 nano: 23.7 (#173), Qwen3.6 Max Preview: 41.7 (#53)
| Benchmark | GPT-5.4 nano | Qwen3.6 Max Preview |
|---|---|---|
| Chess Puzzles | 30% | 20% |
| LMArena Hard Prompts | 1381 | 1457 |
| Mystery Game Puzzles | 9% | 19% |
| DTBench | 80.3% | 87.2% |
| LMCA | 36.9% | 42.5% |
| Epoch Capabilities Index | 145.81 | 149.24 |
| ARC-AGI-2 | 5.7% | — |
| SimpleBench | — | 63% |
| Kagi LLM Benchmark | 39.7% | — |
| NYT Connections (extended) | — | 74.1% |
| ARC-AGI-1 | 51.5% | — |
| CritPt | 9.3% | — |
| ForecastBench | 57.3 | — |
Math Qwen3.6 Max Preview leads
GPT-5.4 nano: 40.9 (#88), Qwen3.6 Max Preview: 54.1 (#46)
| Benchmark | GPT-5.4 nano | Qwen3.6 Max Preview |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 87.8% | 91.1% |
| LMArena Math | 1406 | 1465 |
| FrontierMath (Feb 2025 set) | 25.9% | 23.1% |
| FrontierMath Tier 4 (v1) | 6.3% | 4.2% |
| FrontierMath (Tiers 1-3) | 44.9% | — |
| FrontierMath Tier 4 | 12.2% | — |
| ProofBench | 5% | — |
Knowledge Qwen3.6 Max Preview leads
GPT-5.4 nano: 41.9 (#103), Qwen3.6 Max Preview: 57.6 (#39)
| Benchmark | GPT-5.4 nano | Qwen3.6 Max Preview |
|---|---|---|
| GPQA Diamond | 78.5% | 87.4% |
| SimpleQA Verified | 11.7% | 52% |
| LMArena Expert | 1396 | 1478 |
| Vectara Hallucination Rate | 3.1% | — |
Multimodal Not comparable
GPT-5.4 nano: 36.7 (#78), Qwen3.6 Max Preview: —
| Benchmark | GPT-5.4 nano | Qwen3.6 Max Preview |
|---|---|---|
| LMArena Vision | 1196 | — |
Multilingual Qwen3.6 Max Preview leads
GPT-5.4 nano: 48.6 (#140), Qwen3.6 Max Preview: 54.2 (#48)
| Benchmark | GPT-5.4 nano | Qwen3.6 Max Preview |
|---|---|---|
| LMArena Non-English | 1359 | 1437 |
| LMArena Chinese | 1392 | 1487 |
| LMArena French | 1396 | 1449 |
| LMArena Russian | 1363 | 1445 |
| LMArena Spanish | 1371 | 1454 |
| LMArena German | 1367 | — |
| LMArena Japanese | 1343 | — |
| LMArena Korean | 1320 | — |
Instruction Following Qwen3.6 Max Preview leads
GPT-5.4 nano: 71.9 (#144), Qwen3.6 Max Preview: 75.7 (#55)
| Benchmark | GPT-5.4 nano | Qwen3.6 Max Preview |
|---|---|---|
| LMArena Instruction Following | 1362 | 1438 |
Long Context Qwen3.6 Max Preview leads
GPT-5.4 nano: 41.6 (#137), Qwen3.6 Max Preview: 44.6 (#61)
| Benchmark | GPT-5.4 nano | Qwen3.6 Max Preview |
|---|---|---|
| LMArena Longer Query | 1366 | 1457 |
Writing & Preference Qwen3.6 Max Preview leads
GPT-5.4 nano: 55.7 (#142), Qwen3.6 Max Preview: 63.8 (#60)
| Benchmark | GPT-5.4 nano | Qwen3.6 Max Preview |
|---|---|---|
| LMArena Text | 1372 | 1447 |
| LMArena Creative Writing | 1314 | 1435 |
| LMArena Multi-Turn | 1382 | 1456 |
Frequently asked questions
Is GPT-5.4 nano better than Qwen3.6 Max Preview?
Qwen3.6 Max Preview is the stronger model overall, scoring 51.5 to 41.9 on the Noometry Index. GPT-5.4 nano costs 6.3× less per token, which makes it the better buy when Qwen3.6 Max Preview's lead doesn't matter for your workload.
Which is cheaper, GPT-5.4 nano or Qwen3.6 Max Preview?
GPT-5.4 nano is cheaper. It lists at $0.20 per million input tokens and $1.25 per million output tokens; Qwen3.6 Max Preview lists at $1.30 and $7.80.
Is GPT-5.4 nano or Qwen3.6 Max Preview better for coding?
Qwen3.6 Max Preview scores higher on coding benchmarks: 48.7 versus 43.6 in the Noometry coding category.
Which has the bigger context window?
GPT-5.4 nano does, with 400K tokens against 262K.
How many benchmarks do GPT-5.4 nano and Qwen3.6 Max Preview share?
24 benchmarks have published results for both models. GPT-5.4 nano has 40 scored results on Noometry and Qwen3.6 Max Preview has 29.