Model comparison
GPT-5.4 mini vs Qwen3.5 397B
Head-to-head evidence from 23 shared benchmark results across 6 categories. Overall scores shown here use the public BenchAlign v5 ranking lane.
Public leaderboard positions: GPT-5.4 mini #75 (Estimated); Qwen3.5 397B #71 (Estimated). Intervals and evidence labels describe ranking uncertainty, not a guarantee for a specific workload.
Evidence parity. GPT-5.4 mini and Qwen3.5 397B share 23 comparable benchmark results. 4 of 8 categories are comparable. 7 results are unique to GPT-5.4 mini; 32 to Qwen3.5 397B.
Updated July 23, 2026- Shared results
- 23
- GPT-5.4 mini only
- 7
- Qwen3.5 397B only
- 32
- Comparable categories
- 4 / 8
Pick Qwen3.5 397B if you want the stronger benchmark profile. GPT-5.4 mini only becomes the better choice if agentic is the priority or you need the larger 400K context window.
Confidence note. This is a partial-evidence comparison with 23 shared benchmark results across 6 evidence categories; 4 of 8 categories currently have scoreable aggregates for both models. Treat the verdict as directional until coverage is more balanced.
Why this result
Qwen3.5 397B has the cleaner BenchAlign overall profile here, landing at 57.01 versus 56.77. It is a real lead, but still close enough that category-level strengths matter more than the headline number.
Qwen3.5 397B's sharpest advantage is in mathematics, where it averages 90.6 against 21.7. The single biggest benchmark swing on the page is HLE, 41.5% to 28.7%. GPT-5.4 mini does hit back in agentic, so the answer changes if that is the part of the workload you care about most.
GPT-5.4 mini is also the more expensive model on tokens at $0.75 input / $4.50 output per 1M tokens, versus $0.60 input / $3.60 output per 1M tokens for Qwen3.5 397B. GPT-5.4 mini is the reasoning model in the pair, while Qwen3.5 397B is not. That usually helps on harder chain-of-thought-heavy tests, but it can also mean more latency and more token spend in real use. GPT-5.4 mini gives you the larger context window at 400K, compared with 128K for Qwen3.5 397B.
Category breakdown
Exact category averages are shown below. Not measured means BenchLM does not have enough sourced public coverage for that model and category.
| Category | GPT-5.4 mini | Δ | Qwen3.5 397B |
|---|---|---|---|
| Math | GPT-5.4 mini21.7 | Margin→ 68.9 | Qwen3.5 397B90.6 |
| Agentic | GPT-5.4 mini65.7 | Margin← 9.2 | Qwen3.5 397B56.5 |
| Knowledge | GPT-5.4 mini47.8 | Margin→ 8.8 | Qwen3.5 397B56.6 |
| Multimodal | GPT-5.4 mini76.6 | Margin→ 3.0 | Qwen3.5 397B79.6 |
| Coding | GPT-5.4 miniNot measured | MarginNo overlap | Qwen3.5 397B66.5 |
| Reasoning | GPT-5.4 miniNot measured | MarginNo overlap | Qwen3.5 397B63.2 |
| Multilingual | GPT-5.4 miniNot measured | MarginNo overlap | Qwen3.5 397B84.7 |
| Inst. Following | GPT-5.4 miniNot measured | MarginNo overlap | Qwen3.5 397B92.6 |
Decisive benchmark drivers
The largest measured benchmark gaps in this matchup, with exact reported values.
More
- Source ↗
HLE
KnowledgeA 41.5%B 28.7%Winner: GPT-5.4 miniΔ 12.8HLE: GPT-5.4 mini scored 41.5%; Qwen3.5 397B scored 28.7%. GPT-5.4 mini wins this benchmark. - Source ↗
Terminal-Bench 2.0
AgenticA 60%B 52.5%Winner: GPT-5.4 miniΔ 7.5Terminal-Bench 2.0: GPT-5.4 mini scored 60%; Qwen3.5 397B scored 52.5%. GPT-5.4 mini wins this benchmark. - Source ↗
MMMU-Pro
MultimodalA 76.6%B 79%Winner: Qwen3.5 397BΔ 2.4MMMU-Pro: GPT-5.4 mini scored 76.6%; Qwen3.5 397B scored 79%. Qwen3.5 397B wins this benchmark. - Source ↗
GPQA
KnowledgeA 88%B 88.4%Winner: Qwen3.5 397BΔ 0.4GPQA: GPT-5.4 mini scored 88%; Qwen3.5 397B scored 88.4%. Qwen3.5 397B wins this benchmark.
Operational comparison
Runtime and commercial metrics are compared only when both models have a complete sourced value.
| Metric | GPT-5.4 mini | Qwen3.5 397B | Comparison |
|---|---|---|---|
| Input / output priceUSD per 1M tokens | GPT-5.4 mini$0.75 input / $4.5 output | Qwen3.5 397B$0.6 input / $3.6 output | Qwen3.5 397B has the lower combined listed price. |
| Generation speedtokens per second | GPT-5.4 mini201 tok/s | Qwen3.5 397B96 tok/s | GPT-5.4 mini has the higher measured throughput. |
| First-answer latencyseconds to first token | GPT-5.4 mini3.85 s | Qwen3.5 397B2.44 s | Qwen3.5 397B reaches the first token sooner. |
| Context windowmaximum listed tokens | GPT-5.4 mini400K | Qwen3.5 397B128K | GPT-5.4 mini lists the larger context window. |
Benchmark Deep Dive
AgenticGPT-5.4 mini wins19 benchmarks
| Benchmark | GPT-5.4 mini | Qwen3.5 397B | Result |
|---|---|---|---|
| Terminal-Bench 2.0Source | 60% | 52.5% | GPT-5.4 mini leads |
| OSWorld-VerifiedSource | 72.1% | — | Not comparable |
| MCP AtlasSource | 57.7% | 46.1% | GPT-5.4 mini leads |
| ToolathlonSource | 42.9% | 36.3% | GPT-5.4 mini leads |
| τ²-bench resultsSource | 83.3% | 95.6% | Qwen3.5 397B leads |
| AA Agentic IndexSource | 30.2% | 19.9% | GPT-5.4 mini leads |
| APEX-Agents-AASource | 28.2% | 15.3% | GPT-5.4 mini leads |
| GDPval-AASource | 33.6% | 23.1% | GPT-5.4 mini leads |
| GDPval-AASource | 1171 | 962 | GPT-5.4 mini leads |
| BrowseCompSource | — | 62% | Not comparable |
| Claw-EvalSource | — | 56.8% | Not comparable |
| QwenClawBenchSource | — | 51.8% | Not comparable |
| τ³-bench resultsSource | — | 68.4% | Not comparable |
| VITA-BenchSource | — | 43.7% | Not comparable |
| DeepPlanningSource | — | 37.6% | Not comparable |
| MCP-TasksSource | — | 74.2% | Not comparable |
| WideResearchSource | — | 74.0% | Not comparable |
| Gert LabsSource | — | 46.76% | Not comparable |
| ResearchClawBenchSource | — | 14.2% | Not comparable |
Coding7 benchmarks
| Benchmark | GPT-5.4 mini | Qwen3.5 397B | Result |
|---|---|---|---|
| Vibe Code BenchSource | 47.97% | — | Not comparable |
| AA Coding IndexSource | 56.1% | 48.2% | GPT-5.4 mini leads |
| AA-SciCodeSource | 49.9% | 42.0% | GPT-5.4 mini leads |
| FrontierCode 1.1 MainSource | 27.0% | — | Not comparable |
| SWE-bench VerifiedSource | — | 76.2% | Not comparable |
| LiveCodeBench v6Source | — | 83.6% | Not comparable |
| SWE-bench ProSource | — | 50.9% | Not comparable |
Reasoning4 benchmarks
KnowledgeQwen3.5 397B wins13 benchmarks
| Benchmark | GPT-5.4 mini | Qwen3.5 397B | Result |
|---|---|---|---|
| GPQASource | 88% | 88.4% | Qwen3.5 397B leads |
| HLESource | 41.5% | 28.7% | GPT-5.4 mini leads |
| HLE w/o toolsSource | 28.2% | — | Not comparable |
| Artificial Analysis Intelligence IndexSource | 40.0% | 33.7% | GPT-5.4 mini leads |
| AA-GPQA DiamondSource | 87.5% | 89.3% | Qwen3.5 397B leads |
| AA-HLESource | 26.6% | 27.3% | Qwen3.5 397B leads |
| AA-Omniscience IndexSource | -18.7% | -29.8% | GPT-5.4 mini leads |
| AA-Omniscience AccuracySource | 37.5% | 31.4% | GPT-5.4 mini leads |
| AA-Omniscience Hallucination RateSource | 89.8% | 89.1% | Qwen3.5 397B leads |
| SuperGPQASource | — | 70.4% | Not comparable |
| MMLU-ProSource | — | 87.8% | Not comparable |
| MMLU-ReduxSource | — | 94.9% | Not comparable |
| C-EvalSource | — | 93% | Not comparable |
MathQwen3.5 397B wins7 benchmarks
| Benchmark | GPT-5.4 mini | Qwen3.5 397B | Result |
|---|---|---|---|
| FrontierMath v2 (Tiers 1-3)Source | 28.280% | — | Not comparable |
| FrontierMath v2 (Tier 4)Source | 2.080% | — | Not comparable |
| AIME26Source | — | 93.3% | Not comparable |
| HMMT Feb 2025Source | — | 94.8% | Not comparable |
| HMMT Nov 2025Source | — | 92.7% | Not comparable |
| HMMT Feb 2026Source | — | 87.9% | Not comparable |
| MMAnswerBenchSource | — | 80.9% | Not comparable |
Multilingual2 benchmarks
MultimodalQwen3.5 397B wins8 benchmarks
| Benchmark | GPT-5.4 mini | Qwen3.5 397B | Result |
|---|---|---|---|
| MMMU-ProSource | 76.6% | 79% | Qwen3.5 397B leads |
| MMMU-Pro w/ PythonSource | 78% | — | Not comparable |
| AA-MMMU-ProSource | 73.3% | 77.3% | Qwen3.5 397B leads |
| MathVisionSource | — | 88.6% | Not comparable |
| CharXivSource | — | 80.8% | Not comparable |
| VideoMMMUSource | — | 84.7% | Not comparable |
| ScreenSpot ProSource | — | 65.6% | Not comparable |
| V*Source | — | 95.8% | Not comparable |
Frequently Asked Questions (5)
Which is better, GPT-5.4 mini or Qwen3.5 397B?
Qwen3.5 397B is ahead on BenchLM's BenchAlign leaderboard, 57.01 to 56.77. The biggest single separator in this matchup is HLE, where the scores are 41.5% and 28.7%.
Which is better for knowledge tasks, GPT-5.4 mini or Qwen3.5 397B?
Qwen3.5 397B has the edge for knowledge tasks in this comparison, averaging 56.6 versus 47.8. Inside this category, HLE is the benchmark that creates the most daylight between them.
Which is better for math, GPT-5.4 mini or Qwen3.5 397B?
Qwen3.5 397B has the edge for math in this comparison, averaging 90.6 versus 21.7. GPT-5.4 mini stays close enough that the answer can still flip depending on your workload.
Which is better for agentic tasks, GPT-5.4 mini or Qwen3.5 397B?
GPT-5.4 mini has the edge for agentic tasks in this comparison, averaging 65.7 versus 56.5. Inside this category, GDPval-AA is the benchmark that creates the most daylight between them.
Which is better for multimodal and grounded tasks, GPT-5.4 mini or Qwen3.5 397B?
Qwen3.5 397B has the edge for multimodal and grounded tasks in this comparison, averaging 79.6 versus 76.6. Inside this category, AA-MMMU-Pro is the benchmark that creates the most daylight between them.
Related Comparisons
Explore More
Choose a model with this week’s evidence
Join 2,000+ readers for ranking moves, pricing changes, and the claims that still need proof.
One email each week. Unsubscribe anytime.