Model comparison
GPT-5 Nano vs Qwen3.8 Max
Qwen3.8 Max is the stronger model overall, scoring 56.8 to 33.5 on the Noometry Index. GPT-5 Nano costs 22× less per token, which makes it the better buy when Qwen3.8 Max's lead doesn't matter for your workload.
Last verified . 28 shared benchmarks.
Alibaba (Qwen)Summary
- They share 28 benchmarks with published results for both. GPT-5 Nano scores higher in 0 categories and Qwen3.8 Max in 10 categories; 10 gaps are clear of the uncertainty.
- The widest gap is in math, where Qwen3.8 Max leads 73.2 to 29.4.
- The biggest single-benchmark swing is FrontierMath (Tiers 1-3): 20% for GPT-5 Nano and 74.7% for Qwen3.8 Max.
- GPT-5 Nano is cheaper at $0.05 / $0.40 per million input/output tokens, against $2 / $6 for Qwen3.8 Max.
- Qwen3.8 Max accepts more context: 1M tokens versus 400K.
Side by side
| GPT-5 Nano | Qwen3.8 Max | |
|---|---|---|
| Provider | OpenAI | Alibaba (Qwen) |
| Noometry Index | 33.5 | 56.8 |
| Released | 2025-08-07 | 2026-08-02 |
| Weights | Proprietary | Proprietary |
| Context window | 400K | 1M |
| Max output | 128K | 131K |
| Input $ / M tokens | $0.05 | $2 |
| Output $ / M tokens | $0.40 | $6 |
| Results tracked | 49 | 39 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Qwen3.8 Max leads
GPT-5 Nano: 33.6 (#254), Qwen3.8 Max: 53.5 (#29)
| Benchmark | GPT-5 Nano | Qwen3.8 Max |
|---|---|---|
| LMArena Coding | 1351 | 1502 |
| DeepSWE | — | 57.5% |
| SWE-bench Verified (bash only) | 34.8% | — |
| LMArena WebDev | — | 1674 |
| FrontierSWE | — | 17.8% |
| SciCode | — | 53.2% |
| WeirdML | 38.1% | — |
| ALE-Bench | 718.67 | — |
Agentic & Tool Use Qwen3.8 Max leads
GPT-5 Nano: 25.8 (#106), Qwen3.8 Max: 45.4 (#14)
| Benchmark | GPT-5 Nano | Qwen3.8 Max |
|---|---|---|
| Terminal-Bench | 21.8% | — |
| APEX-Agents | — | 63.3% |
| Berkeley Function Calling Leaderboard | 51.5% | — |
| τ²-bench Banking | — | 55.1% |
| GDP.pdf | — | 23.2% |
Reasoning Qwen3.8 Max leads
GPT-5 Nano: 16.3 (#306), Qwen3.8 Max: 54.4 (#26)
| Benchmark | GPT-5 Nano | Qwen3.8 Max |
|---|---|---|
| Chess Puzzles | 27% | 40% |
| LMArena Hard Prompts | 1328 | 1496 |
| Mystery Game Puzzles | 9% | 38% |
| DTBench | 62.7% | 92% |
| LMCA | 7.9% | 46.2% |
| Epoch Capabilities Index | 139.38 | 156.41 |
| ARC-AGI-2 | 2.6% | — |
| Kagi LLM Benchmark | 62.2% | — |
| NYT Connections (extended) | — | 88.3% |
| ARC-AGI-1 | 20.7% | — |
| CritPt | — | 20% |
| ForecastBench | 59.1 | — |
Math Qwen3.8 Max leads
GPT-5 Nano: 29.4 (#241), Qwen3.8 Max: 73.2 (#20)
| Benchmark | GPT-5 Nano | Qwen3.8 Max |
|---|---|---|
| FrontierMath (Tiers 1-3) | 20% | 74.7% |
| FrontierMath Tier 4 | 2.4% | 46.3% |
| OTIS Mock AIME 2024-2025 | 81.1% | 100% |
| ProofBench | 12% | 58% |
| LMArena Math | 1317 | 1499 |
| Omni-MATH | 54.6% | — |
| MATH Level 5 | 95.2% | — |
| FrontierMath (Feb 2025 set) | 8.3% | — |
| FrontierMath Tier 4 (v1) | 2.1% | — |
Knowledge Qwen3.8 Max leads
GPT-5 Nano: 35.9 (#178), Qwen3.8 Max: 61.7 (#27)
| Benchmark | GPT-5 Nano | Qwen3.8 Max |
|---|---|---|
| GPQA Diamond | 69.4% | 92.7% |
| SimpleQA Verified | 11.7% | 47.3% |
| LMArena Expert | 1321 | 1507 |
| MMLU-Pro | 77.8% | — |
| Vectara Hallucination Rate | 10.5% | — |
| GPQA (HELM) | 67.9% | — |
Multimodal Qwen3.8 Max leads
GPT-5 Nano: 31.3 (#108), Qwen3.8 Max: 37.2 (#75)
| Benchmark | GPT-5 Nano | Qwen3.8 Max |
|---|---|---|
| LMArena Vision | 1159 | 1314 |
| VPCT | 37.2% | — |
| Furniture Assembly | — | 20% |
Multilingual Qwen3.8 Max leads
GPT-5 Nano: 45.3 (#172), Qwen3.8 Max: 56.7 (#18)
| Benchmark | GPT-5 Nano | Qwen3.8 Max |
|---|---|---|
| LMArena Non-English | 1313 | 1472 |
| LMArena Chinese | 1356 | 1538 |
| LMArena German | 1327 | 1483 |
| LMArena Japanese | 1226 | 1467 |
| LMArena Korean | 1269 | 1461 |
| LMArena Russian | 1296 | 1481 |
| LMArena Spanish | 1360 | 1492 |
| LMArena French | — | 1503 |
Instruction Following Qwen3.8 Max leads
GPT-5 Nano: 75.0 (#79), Qwen3.8 Max: 77.6 (#17)
| Benchmark | GPT-5 Nano | Qwen3.8 Max |
|---|---|---|
| LMArena Instruction Following | 1306 | 1479 |
| IFEval | 93.2% | — |
Long Context Qwen3.8 Max leads
GPT-5 Nano: 31.3 (#281), Qwen3.8 Max: 45.6 (#31)
| Benchmark | GPT-5 Nano | Qwen3.8 Max |
|---|---|---|
| LMArena Longer Query | 1312 | 1489 |
| Fiction.LiveBench | 44.4% | — |
Writing & Preference Qwen3.8 Max leads
GPT-5 Nano: 39.1 (#249), Qwen3.8 Max: 67.1 (#30)
| Benchmark | GPT-5 Nano | Qwen3.8 Max |
|---|---|---|
| LMArena Text | 1320 | 1483 |
| LMArena Creative Writing | 1249 | 1479 |
| LMArena Multi-Turn | 1311 | 1489 |
| EQ-Bench Creative Writing | 705 | — |
| WildBench | 80.6% | — |
Frequently asked questions
Is GPT-5 Nano better than Qwen3.8 Max?
Qwen3.8 Max is the stronger model overall, scoring 56.8 to 33.5 on the Noometry Index. GPT-5 Nano costs 22× less per token, which makes it the better buy when Qwen3.8 Max's lead doesn't matter for your workload.
Which is cheaper, GPT-5 Nano or Qwen3.8 Max?
GPT-5 Nano is cheaper. It lists at $0.05 per million input tokens and $0.40 per million output tokens; Qwen3.8 Max lists at $2 and $6.
Is GPT-5 Nano or Qwen3.8 Max better for coding?
Qwen3.8 Max scores higher on coding benchmarks: 53.5 versus 33.6 in the Noometry coding category.
Which has the bigger context window?
Qwen3.8 Max does, with 1M tokens against 400K.
How many benchmarks do GPT-5 Nano and Qwen3.8 Max share?
28 benchmarks have published results for both models. GPT-5 Nano has 49 scored results on Noometry and Qwen3.8 Max has 39.