Model comparison
GPT-5.4 Pro vs Qwen3.8 27B
GPT-5.4 Pro is the stronger model overall, scoring 58.8 to 45.3 on the Noometry Index. Qwen3.8 27B costs 61× less per token, which makes it the better buy when GPT-5.4 Pro's lead doesn't matter for your workload.
Last verified . 2 shared benchmarks.
Alibaba (Qwen)Summary
- They share 2 benchmarks with published results for both. GPT-5.4 Pro scores higher in 3 categories and Qwen3.8 27B in 1 category; 4 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where GPT-5.4 Pro leads 71.0 to 31.5.
- The biggest single-benchmark swing is CritPt: 30% for GPT-5.4 Pro and 5.4% for Qwen3.8 27B.
- Qwen3.8 27B is cheaper at $0.99 / $1.49 per million input/output tokens, against $30 / $180 for GPT-5.4 Pro.
- GPT-5.4 Pro accepts more context: 1.05M tokens versus 262K.
- Qwen3.8 27B has downloadable open weights; the other is API-only.
Side by side
| GPT-5.4 Pro | Qwen3.8 27B | |
|---|---|---|
| Provider | OpenAI | Alibaba (Qwen) |
| Noometry Index | 58.8 | 45.3 |
| Released | 2026-03-05 | 2026-08-14 |
| Weights | Proprietary | Open |
| Context window | 1.05M | 262K |
| Max output | 128K | 33K |
| Input $ / M tokens | $30 | $0.99 |
| Output $ / M tokens | $180 | $1.49 |
| Results tracked | 16 | 28 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Qwen3.8 27B leads
GPT-5.4 Pro: 42.6 (#89), Qwen3.8 27B: 50.3 (#42)
| Benchmark | GPT-5.4 Pro | Qwen3.8 27B |
|---|---|---|
| LMArena WebDev | — | 1592 |
| SciCode | — | 46.6% |
| WeirdML | 57.4% | — |
| LMArena Coding | — | 1487 |
Reasoning GPT-5.4 Pro leads
GPT-5.4 Pro: 71.0 (#10), Qwen3.8 27B: 31.5 (#71)
| Benchmark | GPT-5.4 Pro | Qwen3.8 27B |
|---|---|---|
| CritPt | 30% | 5.4% |
| Epoch Capabilities Index | 159.12 | 149.38 |
| ARC-AGI-2 | 83.3% | — |
| SimpleBench | 74.1% | — |
| NYT Connections (extended) | — | 54.5% |
| ARC-AGI-1 | 94.5% | — |
| Chess Puzzles | 58.6% | — |
| EnigmaEval | 23.8% | — |
| LMArena Hard Prompts | — | 1459 |
| DTBench | — | 88% |
| LMCA | — | 41.4% |
| Surface Evolver Bench | — | 45% |
Math GPT-5.4 Pro leads
GPT-5.4 Pro: 72.0 (#17), Qwen3.8 27B: 35.7 (#139)
| Benchmark | GPT-5.4 Pro | Qwen3.8 27B |
|---|---|---|
| FrontierMath (Tiers 1-3) | 82.5% | — |
| FrontierMath Tier 4 | 58.5% | — |
| ProofBench | — | 16% |
| LMArena Math | — | 1453 |
| FrontierMath (Feb 2025 set) | 50% | — |
| FrontierMath Tier 4 (v1) | 37.5% | — |
Knowledge GPT-5.4 Pro leads
GPT-5.4 Pro: 68.3 (#6), Qwen3.8 27B: 40.9 (#105)
| Benchmark | GPT-5.4 Pro | Qwen3.8 27B |
|---|---|---|
| GPQA Diamond | 94.6% | — |
| Humanity's Last Exam | 44.3% | — |
| SimpleQA Verified | 46.3% | — |
| Vectara Hallucination Rate | 8.3% | — |
| LMArena Expert | — | 1482 |
Multimodal Not comparable
GPT-5.4 Pro: —, Qwen3.8 27B: 40.5 (#36)
| Benchmark | GPT-5.4 Pro | Qwen3.8 27B |
|---|---|---|
| LMArena Vision | — | 1270 |
Multilingual Not comparable
GPT-5.4 Pro: —, Qwen3.8 27B: 54.0 (#54)
| Benchmark | GPT-5.4 Pro | Qwen3.8 27B |
|---|---|---|
| LMArena Non-English | — | 1431 |
| LMArena Chinese | — | 1504 |
| LMArena French | — | 1468 |
| LMArena German | — | 1435 |
| LMArena Japanese | — | 1392 |
| LMArena Korean | — | 1389 |
| LMArena Russian | — | 1421 |
| LMArena Spanish | — | 1444 |
Instruction Following Not comparable
GPT-5.4 Pro: —, Qwen3.8 27B: 75.8 (#54)
| Benchmark | GPT-5.4 Pro | Qwen3.8 27B |
|---|---|---|
| LMArena Instruction Following | — | 1436 |
Long Context Not comparable
GPT-5.4 Pro: —, Qwen3.8 27B: 44.3 (#66)
| Benchmark | GPT-5.4 Pro | Qwen3.8 27B |
|---|---|---|
| LMArena Longer Query | — | 1448 |
Writing & Preference Not comparable
GPT-5.4 Pro: —, Qwen3.8 27B: 65.9 (#41)
| Benchmark | GPT-5.4 Pro | Qwen3.8 27B |
|---|---|---|
| LMArena Text | — | 1441 |
| LMArena Creative Writing | — | 1382 |
| EQ-Bench Creative Writing | — | 1671 |
| LMArena Multi-Turn | — | 1443 |
Frequently asked questions
Is GPT-5.4 Pro better than Qwen3.8 27B?
GPT-5.4 Pro is the stronger model overall, scoring 58.8 to 45.3 on the Noometry Index. Qwen3.8 27B costs 61× less per token, which makes it the better buy when GPT-5.4 Pro's lead doesn't matter for your workload.
Which is cheaper, GPT-5.4 Pro or Qwen3.8 27B?
Qwen3.8 27B is cheaper. It lists at $0.99 per million input tokens and $1.49 per million output tokens; GPT-5.4 Pro lists at $30 and $180.
Is GPT-5.4 Pro or Qwen3.8 27B better for coding?
Qwen3.8 27B scores higher on coding benchmarks: 50.3 versus 42.6 in the Noometry coding category.
Which has the bigger context window?
GPT-5.4 Pro does, with 1.05M tokens against 262K.
How many benchmarks do GPT-5.4 Pro and Qwen3.8 27B share?
2 benchmarks have published results for both models. GPT-5.4 Pro has 16 scored results on Noometry and Qwen3.8 27B has 28.