Qwen 3.6 Plus vs GPT 5.4 Mini: Benchmark Comparison
Qwen 3.6 Plus has the higher score on 9 of 16 shared benchmarks; GPT 5.4 Mini leads on 6.
The largest observed score gap is 22.40 pts on Vibe Code Bench v1.1 , where GPT 5.4 Mini leads.
Reported ±1 standard-error ranges overlap on 6 of 16 shared benchmarks with comparable uncertainty data. This is not a pairwise statistical significance test.
Shared benchmark results
Scores and reported standard errors come from the latest visible Vals benchmark version available for both models.
| Benchmark | Qwen 3.6 Plus | GPT 5.4 Mini | Gap | Reported uncertainty |
|---|---|---|---|---|
| Harvey's Legal Agent Benchmark | 1.25% ±0.83 | 0.00% ±0.00 | 1.25 pts | Reported ±1 SE ranges do not overlap |
| Legal Research Bench | 14.90% ±2.48 | 12.50% ±2.30 | 2.40 pts | Reported ±1 SE ranges overlap |
| EMB | 32.87% ±2.81 | 45.43% ±3.65 | 12.56 pts | Reported ±1 SE ranges do not overlap |
| Finance Agent (v2) | 40.85% ±0.13 | 45.36% ±0.45 | 4.51 pts | Reported ±1 SE ranges do not overlap |
| MortgageTax | 67.97% ±0.91 | 63.51% ±0.91 | 4.45 pts | Reported ±1 SE ranges do not overlap |
| Tax Agent Bench | 32.84% ±2.75 | 36.59% ±3.10 | 3.74 pts | Reported ±1 SE ranges overlap |
| TaxEval v2 | 74.73% ±0.85 | 71.22% ±0.90 | 3.52 pts | Reported ±1 SE ranges do not overlap |
| GPQA Diamond | 87.37% ±1.67 | 83.08% ±2.46 | 4.29 pts | Reported ±1 SE ranges do not overlap |
| MMLU Pro | 87.67% ±0.32 | 84.55% ±0.36 | 3.11 pts | Reported ±1 SE ranges do not overlap |
| MMMU Pro | 84.16% ±0.88 | 79.25% ±0.97 | 4.91 pts | Reported ±1 SE ranges do not overlap |
| SAGE | 44.86% ±3.43 | 50.81% ±3.40 | 5.95 pts | Reported ±1 SE ranges overlap |
| Code Migration | 11.10% ±1.31 | 12.94% ±3.64 | 1.83 pts | Reported ±1 SE ranges overlap |
| LiveCodeBench | 85.95% ±0.98 | 81.47% ±1.09 | 4.49 pts | Reported ±1 SE ranges do not overlap |
| ProgramBench | 0.00% ±0.00 | 0.00% ±0.00 | 0.00 pts | Reported ±1 SE ranges overlap |
| SWE-bench | 73.40% ±1.98 | 73.00% ±1.99 | 0.40 pts | Reported ±1 SE ranges overlap |
| Vibe Code Bench v1.1 | 25.57% ±3.91 | 47.97% ±5.61 | 22.40 pts | Reported ±1 SE ranges do not overlap |
Performance by category
| Category | Qwen 3.6 Plus average | GPT 5.4 Mini average |
|---|---|---|
| Legal | 8.08% | 6.25% |
| Finance | 49.85% | 52.42% |
| Academic | 86.40% | 82.29% |
| Education | 44.86% | 50.81% |
| Coding | 39.20% | 43.07% |
Cost and latency
Cost per test appears only where the benchmark reports it for both models. Latency is the measured completion time for that benchmark.
| Benchmark | Qwen 3.6 Plus cost | GPT 5.4 Mini cost | Qwen 3.6 Plus latency | GPT 5.4 Mini latency |
|---|---|---|---|---|
| Harvey's Legal Agent Benchmark | $3.94 | $1.00 | 14m48s | 11m21s |
| Legal Research Bench | $0.70 | $1.86 | 10m59s | 37m36s |
| EMB | $2.87 | $2.15 | 31m14s | 43m21s |
| Finance Agent (v2) | $0.81 | $1.20 | 10m12s | 35m54s |
| MortgageTax | N/A | N/A | 78.18s | 2m06s |
| Tax Agent Bench | $0.30 | $0.81 | 6m18s | 16m46s |
| TaxEval v2 | N/A | N/A | 70.63s | 44.58s |
| GPQA Diamond | N/A | N/A | 2m02s | 63.11s |
| MMLU Pro | N/A | N/A | 46.18s | 20.00s |
| MMMU Pro | N/A | N/A | 111.94s | 50.81s |
| SAGE | N/A | N/A | 3m49s | 2m21s |
| Code Migration | $4.78 | $1.95 | 29m50s | 28m53s |
| LiveCodeBench | N/A | N/A | 7m56s | 3m09s |
| ProgramBench | N/A | N/A | 33m01s | 53m45s |
| SWE-bench | $0.87 | $0.51 | 7m04s | 5m26s |
| Vibe Code Bench v1.1 | $5.45 | $1.19 | 38m10s | 34m15s |
Results available only for Qwen 3.6 Plus
- LegalBench
- MedCode
- MedScribe
Results available only for GPT 5.4 Mini
- Vals Index
- Terminal-Bench 4.0