Qwen3.6-35B-A3B vs Qwen3.6-27B vs Qwen3.5-9B vs Qwen3.5-27B vs Qwen3.8-Flash-Next vs Qwen3.5-35B-A3B vs Qwen3.8-27B
New ComparisonScores are fetched live from the current database on every visit — nothing is cached.
Winner
Qwen3.8-Flash-Next
Qwen3.8-Flash-Next wins this comparison with an overall score of 58 vs 52 for Qwen3.5-27B, leading 3 of 4 shared benchmarks.
Qwen3.8-Flash-Next
57.7
Qwen3.5-27B
52.4
Qwen3.5-35B-A3B
51.1
Qwen3.6-27B
48.2
Qwen3.6-35B-A3B
44
Qwen3.8-27B
39.6
Qwen3.5-9B
20.4
Based on 4 of 157 benchmarks scored by all 7 selected models.
Benchmark Scores Comparison
Qwen3.6-35B-A3B
Qwen3.6-27B
Qwen3.5-9B
Qwen3.5-27B
Qwen3.8-Flash-Next
Qwen3.5-35B-A3B
Qwen3.8-27B
Qwen3.6-35B-A3B
35.0B params
Score: 100.0
Qwen3.6-27B
27.0B params
Score: 100.0
Qwen3.5-9B
Score: 15.0
Qwen3.5-27B
27.0B params
Score: 100.0
Qwen3.8-Flash-Next
125.0B params
Score: 70.0
Qwen3.5-35B-A3B
35.0B params
Score: 100.0
Qwen3.8-27B
27.0B params
Score: 100.0
Coverage matrix
7 models × 157 benchmarks — hover a benchmark name for its category
Shading = score relative to the best score among the selected models:
Low
Mid
High
Missing (—)
The totals row shows how many of the 157 benchmarks each model has scores for.
| Benchmark | Qwen3.5-27B Qwen Team (Alibaba Cloud) | Qwen3.5-35B-A3B Qwen Team (Alibaba Cloud) | Qwen3.6-35B-A3B Qwen Team (Alibaba Cloud) | Qwen3.6-27B Qwen Team (Alibaba Cloud) | Qwen3.8-27B Qwen Team (Alibaba Cloud) | Qwen3.8-Flash-Next Qwen Team (Alibaba) | Qwen3.5-9B Qwen Team (Alibaba Cloud) |
|---|---|---|---|---|---|---|---|
| HMMT Feb 25 | 92 | 89 | 90.7 | 93.8 | — | — | — |
| DynaMath | 87.7 | 85 | 82.8 | 85.6 | — | — | — |
| GPQA Diamond | 85.5 | 84.2 | 86 | 87.8 | 89.2 | 91.7 | — |
| IFBench (prompt loose) | — | — | — | 69.1 | 79.5 | — | — |
| MathVista (mini) | 87.8 | 86.2 | 86.4 | 87.4 | — | — | — |
| MMBench EN-DEV v1.1 | 92.6 | 91.5 | 92.8 | 92.3 | — | — | — |
| Nuscene | 15.2 | 14.6 | — | — | — | — | — |
| LiveCodeBench v6 | 80.7 | 74.6 | 80.4 | 83.9 | 90.3 | 91.9 | — |
| SWE-bench Pro | 51.2 | 44.6 | 49.5 | 53.5 | 61.7 | 62.5 | 33.8 |
| BFCL-V4 | 68.5 | 67.3 | 64.7 | — | 72.4 | — | 66.1 |
| RealWorldQA | 83.7 | 84.1 | 85.3 | 84.1 | 85.9 | 88.5 | — |
| Agents' Last Exam (Score) | — | — | — | — | 42.9 | 51.2 | — |
| EmbSpatialBench | 84.5 | 83.1 | 84.3 | 84.6 | — | — | — |
| AA-LCR | 66.1 | 58.5 | 69.3 | 73.3 | 80.3 | — | 63 |
| Hypersim | 13 | 13.1 | — | — | — | — | — |
| MathVision | 86 | 83.9 | — | 85.1 | 90 | 90.6 | — |
| AIME 26 | 92.6 | 91 | 92.7 | 94.1 | 98.8 | — | 88.2 |
| BrowseComp-zh | 62.1 | 69.5 | — | — | — | — | — |
| MAXIFE | 88 | 86.6 | — | — | — | — | — |
| AndroidWorld | 64.2 | 71.1 | — | 70.3 | 81.9 | 84.5 | — |
| AgentWorldBench Terminal | — | 46.13 | — | — | — | — | — |
| Agents' Last Exam | — | — | — | 10.6 | 20.4 | 24.3 | — |
| Toolathlon Verified | — | — | 41.7 | — | 67.1 | 73.5 | — |
| RecreationBench | — | — | — | 29.8 | 47.1 | 49.9 | — |
| VITA-Bench | 41.8 | 29.1 | 35.6 | — | — | — | 15.6 |
| TAU3-Bench | 68.4 | 68.9 | 67.2 | — | — | — | 9.3 |
| ScreenSpot Pro | 70.3 | 68.6 | — | 76.1 | — | — | — |
| OJBench | 40.1 | 36 | — | — | — | — | — |
| AgentWorldBench Overall | — | 47.73 | — | — | — | — | — |
| WMDP (Bio) | — | — | — | 84.8 | — | — | — |
| AutomationBench (Public) | — | — | 24.9 | — | 58.7 | — | — |
| MMStar | 81 | 81.9 | 80.7 | 81.4 | — | — | — |
| GDPVal | — | — | 67.4 | — | 63.5 | — | — |
| OCRBench | 89.4 | 91 | — | — | — | — | — |
| Seal-0 | 47.2 | 41.4 | — | — | — | — | — |
| PMC-VQA | 62.4 | 62 | — | — | — | — | — |
| DeepSearch QA | — | — | — | 71.1 | — | — | — |
| MMLongBench-Doc | 60.2 | 59.5 | — | — | — | — | — |
| SWE Atlas - TW | — | 9.8 | 13.3 | — | — | — | 4.4 |
| WildClawBench | — | — | 68.7 | 43.2 | — | — | — |
| AgentWorldBench Web | — | 47.1 | — | — | — | — | — |
| SWE Atlas - QnA | — | 13.2 | 15.5 | — | — | — | 9.2 |
| AI2D_TEST | 92.9 | 92.6 | 92.7 | — | — | — | — |
| LingoQA | 82 | 79.2 | — | — | — | — | — |
| ParseBench Text Formatting | — | — | 58.3 | — | — | — | — |
| NOVA-63 | 58.1 | 57.1 | — | — | — | — | — |
| NL2Repo | 27.3 | 20.5 | 29.4 | 36.2 | 42.3 | — | 16.2 |
| HPCT | — | — | — | 48.7 | — | — | — |
| MMMU-Pro | 75 | 75.1 | 75.3 | 75.8 | — | — | — |
| INCLUDE | 81.6 | 79.7 | — | — | — | — | — |
| QwenClawBench | 52.2 | 47.7 | 52.6 | 53.4 | — | — | — |
| BabyVision | 44.6 | 38.4 | — | 28.9 | 65.7 | — | — |
| SLAKE | 80 | 78.7 | — | — | — | — | — |
| MMMLU | 85.9 | 85.2 | — | — | — | — | — |
| Claw-Eval Avg | 64.3 | 65.4 | 68.7 | 72.4 | 56.9 | 60.4 | 53.2 |
| VideoMMMU | 82.3 | 80.4 | 83.7 | 84.4 | — | — | — |
| HMMT Nov 25 | 89.8 | 89.2 | 89.1 | 90.7 | — | — | — |
| Multi-Challenge | 60.8 | 60 | 58.2 | — | 63.6 | — | — |
| PolyMATH | 71.2 | 64.4 | — | — | — | — | — |
| SWE-MM | — | — | — | 25.7 | 38.6 | — | — |
| MMLU-ProX | 82.2 | 81 | — | — | — | — | — |
| MMLU-Pro | 86.1 | 85.3 | 85.2 | 86.2 | 85.7 | — | — |
| CodeForces | 1899 | 2028 | — | — | — | — | — |
| SuperGPQA | 65.6 | 63.4 | 64.7 | 66 | — | — | — |
| OSWorld-Verified | 56.2 | 54.5 | — | 63.9 | 84.3 | — | — |
| TIR-Bench | 59.8 | 55.5 | — | — | — | — | — |
| Global PIQA | 87.5 | 86.6 | — | — | — | — | — |
| AgentWorldBench Search | — | 25.98 | — | — | — | — | — |
| IFEval | 95 | 91.9 | — | — | — | — | 91.5 |
| HLE (with tools) | 48.5 | 47.4 | 28.9 | — | 30.8 | 35.9 | — |
| LVBench | 73.6 | 71.4 | 71.4 | — | 72.4 | 76.6 | — |
| CharXiv (RQ) | 79.5 | 77.5 | 78 | 78.4 | 83.7 | 84.6 | — |
| GDP.pdf | — | — | 6.6 | — | 11.2 | — | — |
| BrowseComp | 61 | 61 | 62 | — | — | — | 8.3 |
| MedXpertQA-MM | 62.4 | 61.4 | — | — | — | — | — |
| MMLU-Redux | 93.2 | 93.3 | 93.3 | 93.5 | — | — | — |
| NL2Repo-Bench | — | — | — | — | 42.3 | 48.1 | — |
| SkillsBench | — | — | — | 46.6 | — | — | — |
| WebArena-Verified | — | — | — | 48.8 | 64.8 | — | — |
| ParseBench Text Content | — | — | 90.7 | — | — | — | — |
| CoWorkBench | — | — | — | 61 | 70.7 | 73.9 | — |
| Lab Bench (ProtocolQA) | — | — | — | 69.1 | — | — | — |
| MCPMark | 36.3 | 27 | 37 | — | — | — | 13.4 |
| AA-Omniscience Index (public set) | — | — | -7.2 | — | -0.5 | — | — |
| SWE-bench Multilingual | 69.3 | 60.3 | 67.2 | 71.3 | 73.8 | 81 | 43.3 |
| MMVU | 73.3 | 72.3 | — | — | — | — | — |
| FullStackBench zh | 57.4 | 55 | — | — | — | — | — |
| MLVU | 85.9 | 85.6 | 86.2 | 86.6 | — | — | — |
| Beam128K | — | — | — | 63 | — | — | — |
| IFBench | 76.5 | 70.2 | 66.6 | 70.8 | 79.5 | 81.3 | 64.5 |
| FullStackBench en | 60.1 | 58.1 | — | — | — | — | — |
| SWE-bench Verified | 75 | 70 | 73.4 | 77.2 | 74.6 | — | 53.1 |
| IMOAnswerBench | 79.9 | 76.8 | 78.9 | 80.8 | — | — | 69.8 |
| Humanity's Last Exam | 24.3 | 22.4 | 21.4 | 24 | 30.8 | — | 14.3 |
| CountBench | 97.8 | 97.8 | 96.1 | 97.8 | — | — | — |
| WMDP (Chem) | — | — | — | 74.8 | — | — | — |
| ERQA | 60.5 | 64.8 | 61.8 | 62.5 | 65.5 | 72.3 | — |
| VideoMME (w/o sub.) | 82.8 | 82.5 | 82.5 | — | — | — | — |
| Terminal-Bench 2.0 | 41.6 | 40.5 | 51.5 | 59.3 | — | — | — |
| DeepPlanning | 22.6 | 22.8 | 25.9 | — | — | — | — |
| QwenSWEBench | — | — | — | 49.3 | 79 | — | — |
| QwenWebBench | 1068 | 978 | 1397 | 1487 | — | — | — |
| TAU2-Bench | 79 | 81.2 | — | — | — | — | 79.1 |
| ODInW13 | 41.1 | 42.6 | 50.8 | — | — | — | — |
| VlmsAreBlind | 96.9 | 97 | 96.6 | 97 | — | — | — |
| Terminal-Bench 2.1 (Terminus-2) | — | 41.4 | 52.5 | 63.4 | 73 | — | 21.3 |
| DeepSWE | — | — | 0 | — | — | — | — |
| OmniDocBench 1.5 | 88.9 | 89.3 | 89.9 | 89.4 | 91.1 | — | — |
| HallusionBench | 70 | 67.9 | 69.8 | — | — | — | — |
| OCRBench | 89.4 | 89.4 | 90 | 89.4 | — | — | — |
| HMMT Feb 26 | 84.3 | 78.7 | 83.6 | 84.3 | 95.8 | — | 70.8 |
| Terminal-Bench Lite | — | — | 63.3 | — | 67 | — | — |
| AgentWorldBench OS | — | 56.27 | — | — | — | — | — |
| AgentWorldBench MCP | — | 57.87 | — | — | — | — | — |
| GPQA | — | — | — | — | — | — | 77.2 |
| AgentWorldBench SWE | — | 47.58 | — | — | — | — | — |
| CC-OCR | 81 | 80.7 | 81.9 | 81.2 | — | — | — |
| RefCOCO (avg) | 90.9 | 89.2 | 92 | 92.5 | — | — | — |
| JobBench | — | — | — | 21.8 | 33.4 | 55.7 | — |
| Artificial Analysis Intelligence Index | — | — | 32 | — | — | — | — |
| SciCode (subtask) | — | — | — | 39.8 | — | — | — |
| Tool Decathlon | 31.5 | 28.7 | 26.9 | — | — | — | — |
| Frontier-Bench v0.1 | — | — | 1.4 | — | — | — | — |
| OSWorld 2.0 (Partial) | — | — | — | — | 48 | 52.3 | — |
| ParseBench Mean | — | — | 44.1 | — | — | — | — |
| OSWorld 2.0 (Binary) | — | — | — | — | 19.4 | 19.4 | — |
| ZEROBench_sub | 36.2 | 34.1 | 34.4 | — | — | — | — |
| C-Eval | 90.5 | 90.2 | 90 | 91.4 | — | — | — |
| WideSearch | 66.4 | 59.1 | 60.1 | — | — | — | — |
| SimpleVQA | 56 | 58.3 | 58.9 | 56.1 | — | — | — |
| VideoMME (w sub.) | 87 | 86.6 | 86.6 | 87.7 | — | — | — |
| Workspace Bench | — | — | — | — | — | — | 25.5 |
| SkillsBench Avg5 | 27.2 | 4.4 | 28.7 | 48.2 | — | — | — |
| Vision2Web | — | — | — | 45 | 62.9 | 64 | — |
| TauBench V3 Banking | — | — | — | 16.7 | — | — | — |
| Claw-Eval Pass^3 | 46.2 | 51 | 50 | 60.6 | 57.4 | 64.4 | — |
| WMT24++ (en→xx) | 77.6 | 76.3 | — | — | — | — | — |
| LongBench v2 | 60.6 | 59 | — | — | — | — | — |
| MCP-Atlas | 68.4 | 62.4 | 62.8 | 62.5 | — | — | 47.4 |
| SWE Atlas - RF | — | 10.2 | 11.4 | — | — | — | 4.3 |
| V-Star | 93.7 | 92.7 | 90.1 | 94.7 | — | — | — |
| AgentWorldBench Android | — | 53.18 | — | — | — | — | — |
| DeepSWE 1.1 | — | — | — | 13.3 | 42.2 | 58.7 | — |
| Terminal-Bench 2.1 (Claude Code) | — | 38.9 | 49.2 | — | — | — | 18.9 |
| Chatbot Korean (internal) | — | — | 74.5 | — | 90.1 | — | — |
| MBCT | — | — | — | 45.9 | — | — | — |
| VCT | — | — | — | 33.7 | — | — | — |
| Gaokao 2026 | — | — | — | — | — | — | 135.5 |
| Gaia2 | — | — | — | 40 | — | — | — |
| SUNRGBD | 35.4 | 33.4 | — | — | — | — | — |
| SciCode | — | — | 35.6 | — | 44.6 | — | 32.7 |
| RefSpatialBench | 67.7 | 63.5 | 64.3 | 70 | — | — | — |
| BrowseComp Plus | — | — | 34.8 | — | 58 | — | — |
| MMMU | 82.3 | 81.4 | 81.7 | 82.9 | — | — | — |
| GDPVal-AA v2 | — | — | — | 1141 | — | — | — |
| MVBench | 74.6 | 74.8 | 74.6 | 75.5 | — | — | — |
| ZEROBench | 10 | 8 | — | — | — | — | — |
| Covered | 98/157 | 111/157 | 85/157 | 78/157 | 50/157 | 25/157 | 28/157 |
Scorecard
Overall 0–100 score — category strengths (50%), shared-benchmark wins (30%), coverage (20%)
| Model | Overall score | Shared benchmark wins | Coverage | Strongest at | Weakest at |
|---|---|---|---|---|---|
| Qwen3.8-Flash-Next Winner | 57.7 | 3 of 4 | 25 of 157 (16.0%) | General Agents | STEM & Reasoning |
| Qwen3.5-27B | 52.4 | 0 of 4 | 98 of 157 (62.0%) | Video Understanding | Agentic Coding |
| Qwen3.5-35B-A3B | 51.1 | 0 of 4 | 111 of 157 (71.0%) | Video Understanding | Agentic Coding |
| Qwen3.6-27B | 48.2 | 1 of 4 | 78 of 157 (50.0%) | Knowledge | General Agents |
| Qwen3.6-35B-A3B | 44 | 0 of 4 | 85 of 157 (54.0%) | Knowledge | Multilingual |
| Qwen3.8-27B | 39.6 | 0 of 4 | 50 of 157 (32.0%) | Other | Spatial Intelligence |
| Qwen3.5-9B | 20.4 | 0 of 4 | 28 of 157 (18.0%) | Other | Agentic Coding |
Category strengths
Where each model stands, per benchmark category — score relative to the best model, adjusted for benchmark coverage. This also ranks models when they share few benchmarks.
| Category | Qwen3.8-Flash-Next | Qwen3.5-27B | Qwen3.5-35B-A3B | Qwen3.6-27B | Qwen3.6-35B-A3B | Qwen3.8-27B | Qwen3.5-9B |
|---|---|---|---|---|---|---|---|
| General Agents 40 benchmarks | 70 | 68.6 | 62.2 | 31.6 | 57.9 | 68.5 | 31.1 |
| Agentic Coding 25 benchmarks | 64.8 | 56.4 | 53.9 | 62.3 | 73.9 | 71.7 | 19.7 |
| Vision & Language 20 benchmarks | 61.8 | 93.4 | 86.2 | 69.9 | 74.4 | 62 | — |
| STEM & Reasoning 12 benchmarks | 59.1 | 91.6 | 86.9 | 80.7 | 75.2 | 71.8 | 49.4 |
| Document Understanding 10 benchmarks | 61.4 | 92.6 | 91.2 | 71.7 | 80.6 | 66.2 | — |
| Spatial Intelligence 10 benchmarks | 59.5 | 84.8 | 74 | 73 | 73 | 47.5 | — |
| Multilingual 9 benchmarks | — | 95 | 83 | — | 0 | 60 | — |
| Video Understanding 7 benchmarks | 61.4 | 96 | 93 | 80.7 | 88.7 | 54.1 | — |
| Knowledge 5 benchmarks | — | 88.1 | 85.6 | 91 | 92.4 | 72.3 | — |
| Instruction Following 4 benchmarks | 66.3 | 82.4 | 76.6 | 46.9 | 57.9 | 87.8 | 64.1 |
| Long Context 4 benchmarks | — | 77.5 | 71.6 | 63.9 | 53.5 | 85 | 54.9 |
| Other 3 benchmarks | — | — | — | — | 76.7 | 100 | 69.2 |
Fully comparable benchmarks
Scored by all 7 selected models (4)
| Benchmark | Category | Qwen3.6-35B-A3B | Qwen3.6-27B | Qwen3.5-9B | Qwen3.5-27B | Qwen3.8-Flash-Next | Qwen3.5-35B-A3B | Qwen3.8-27B |
|---|---|---|---|---|---|---|---|---|
| SWE-bench Pro | coding_agent | 49.5 | 53.5 | 33.8 | 51.2 | 62.5 | 44.6 | 61.7 |
| Claw-Eval Avg | coding_agent | 68.7 | 72.4 | 53.2 | 64.3 | 60.4 | 65.4 | 56.9 |
| SWE-bench Multilingual | coding_agent | 67.2 | 71.3 | 43.3 | 69.3 | 81 | 60.3 | 73.8 |
| IFBench | instruction_following | 66.6 | 70.8 | 64.5 | 76.5 | 81.3 | 70.2 | 79.5 |
Partial coverage benchmarks
Scored by at least one but not all selected models (153) — missing scores show as —
| Benchmark | Category | Qwen3.6-35B-A3B | Qwen3.6-27B | Qwen3.5-9B | Qwen3.5-27B | Qwen3.8-Flash-Next | Qwen3.5-35B-A3B | Qwen3.8-27B |
|---|---|---|---|---|---|---|---|---|
| HMMT Feb 25 4/7 models | stem_reasoning | 90.7 | 93.8 | — | 92 | — | 89 | — |
| DynaMath 4/7 models | vision_language | 82.8 | 85.6 | — | 87.7 | — | 85 | — |
| GPQA Diamond 6/7 models | stem_reasoning | 86 | 87.8 | — | 85.5 | 91.7 | 84.2 | 89.2 |
| IFBench (prompt loose) 2/7 models | instruction_following | — | 69.1 | — | — | — | — | 79.5 |
| MathVista (mini) 4/7 models | vision_language | 86.4 | 87.4 | — | 87.8 | — | 86.2 | — |
| MMBench EN-DEV v1.1 4/7 models | vision_language | 92.8 | 92.3 | — | 92.6 | — | 91.5 | — |
| Nuscene 2/7 models | spatial_intelligence | — | — | — | 15.2 | — | 14.6 | — |
| LiveCodeBench v6 6/7 models | stem_reasoning | 80.4 | 83.9 | — | 80.7 | 91.9 | 74.6 | 90.3 |
| BFCL-V4 5/7 models | general_agent | 64.7 | — | 66.1 | 68.5 | — | 67.3 | 72.4 |
| RealWorldQA 6/7 models | vision_language | 85.3 | 84.1 | — | 83.7 | 88.5 | 84.1 | 85.9 |
| Agents' Last Exam (Score) 2/7 models | general_agent | — | — | — | — | 51.2 | — | 42.9 |
| EmbSpatialBench 4/7 models | spatial_intelligence | 84.3 | 84.6 | — | 84.5 | — | 83.1 | — |
| AA-LCR 6/7 models | long_context | 69.3 | 73.3 | 63 | 66.1 | — | 58.5 | 80.3 |
| Hypersim 2/7 models | spatial_intelligence | — | — | — | 13 | — | 13.1 | — |
| MathVision 5/7 models | vision_language | — | 85.1 | — | 86 | 90.6 | 83.9 | 90 |
| AIME 26 6/7 models | stem_reasoning | 92.7 | 94.1 | 88.2 | 92.6 | — | 91 | 98.8 |
| BrowseComp-zh 2/7 models | general_agent | — | — | — | 62.1 | — | 69.5 | — |
| MAXIFE 2/7 models | multilingual | — | — | — | 88 | — | 86.6 | — |
| AndroidWorld 5/7 models | general_agent | — | 70.3 | — | 64.2 | 84.5 | 71.1 | 81.9 |
| AgentWorldBench Terminal 1/7 models | general_agent | — | — | — | — | — | 46.13 | — |
| Agents' Last Exam 3/7 models | general_agent | — | 10.6 | — | — | 24.3 | — | 20.4 |
| Toolathlon Verified 3/7 models | general_agent | 41.7 | — | — | — | 73.5 | — | 67.1 |
| RecreationBench 3/7 models | general_agent | — | 29.8 | — | — | 49.9 | — | 47.1 |
| VITA-Bench 4/7 models | general_agent | 35.6 | — | 15.6 | 41.8 | — | 29.1 | — |
| TAU3-Bench 4/7 models | general_agent | 67.2 | — | 9.3 | 68.4 | — | 68.9 | — |
| ScreenSpot Pro 3/7 models | general_agent | — | 76.1 | — | 70.3 | — | 68.6 | — |
| OJBench 2/7 models | stem_reasoning | — | — | — | 40.1 | — | 36 | — |
| AgentWorldBench Overall 1/7 models | general_agent | — | — | — | — | — | 47.73 | — |
| WMDP (Bio) 1/7 models | safety | — | 84.8 | — | — | — | — | — |
| AutomationBench (Public) 2/7 models | general_agent | 24.9 | — | — | — | — | — | 58.7 |
| MMStar 4/7 models | vision_language | 80.7 | 81.4 | — | 81 | — | 81.9 | — |
| GDPVal 2/7 models | general_agent | 67.4 | — | — | — | — | — | 63.5 |
| OCRBench 2/7 models | document_understanding | — | — | — | 89.4 | — | 91 | — |
| Seal-0 2/7 models | general_agent | — | — | — | 47.2 | — | 41.4 | — |
| PMC-VQA 2/7 models | vision_language | — | — | — | 62.4 | — | 62 | — |
| DeepSearch QA 1/7 models | general_agent | — | 71.1 | — | — | — | — | — |
| MMLongBench-Doc 2/7 models | document_understanding | — | — | — | 60.2 | — | 59.5 | — |
| SWE Atlas - TW 3/7 models | coding_agent | 13.3 | — | 4.4 | — | — | 9.8 | — |
| WildClawBench 2/7 models | coding_agent | 68.7 | 43.2 | — | — | — | — | — |
| AgentWorldBench Web 1/7 models | general_agent | — | — | — | — | — | 47.1 | — |
| SWE Atlas - QnA 3/7 models | coding_agent | 15.5 | — | 9.2 | — | — | 13.2 | — |
| AI2D_TEST 3/7 models | document_understanding | 92.7 | — | — | 92.9 | — | 92.6 | — |
| LingoQA 2/7 models | spatial_intelligence | — | — | — | 82 | — | 79.2 | — |
| ParseBench Text Formatting 1/7 models | document_understanding | 58.3 | — | — | — | — | — | — |
| NOVA-63 2/7 models | multilingual | — | — | — | 58.1 | — | 57.1 | — |
| NL2Repo 6/7 models | coding_agent | 29.4 | 36.2 | 16.2 | 27.3 | — | 20.5 | 42.3 |
| HPCT 1/7 models | safety | — | 48.7 | — | — | — | — | — |
| MMMU-Pro 4/7 models | vision_language | 75.3 | 75.8 | — | 75 | — | 75.1 | — |
| INCLUDE 2/7 models | multilingual | — | — | — | 81.6 | — | 79.7 | — |
| QwenClawBench 4/7 models | coding_agent | 52.6 | 53.4 | — | 52.2 | — | 47.7 | — |
| BabyVision 4/7 models | vision_language | — | 28.9 | — | 44.6 | — | 38.4 | 65.7 |
| SLAKE 2/7 models | vision_language | — | — | — | 80 | — | 78.7 | — |
| MMMLU 2/7 models | multilingual | — | — | — | 85.9 | — | 85.2 | — |
| VideoMMMU 4/7 models | video_understanding | 83.7 | 84.4 | — | 82.3 | — | 80.4 | — |
| HMMT Nov 25 4/7 models | stem_reasoning | 89.1 | 90.7 | — | 89.8 | — | 89.2 | — |
| Multi-Challenge 4/7 models | instruction_following | 58.2 | — | — | 60.8 | — | 60 | 63.6 |
| PolyMATH 2/7 models | multilingual | — | — | — | 71.2 | — | 64.4 | — |
| SWE-MM 2/7 models | coding_agent | — | 25.7 | — | — | — | — | 38.6 |
| MMLU-ProX 2/7 models | multilingual | — | — | — | 82.2 | — | 81 | — |
| MMLU-Pro 5/7 models | knowledge | 85.2 | 86.2 | — | 86.1 | — | 85.3 | 85.7 |
| CodeForces 2/7 models | stem_reasoning | — | — | — | 1899 | — | 2028 | — |
| SuperGPQA 4/7 models | knowledge | 64.7 | 66 | — | 65.6 | — | 63.4 | — |
| OSWorld-Verified 4/7 models | general_agent | — | 63.9 | — | 56.2 | — | 54.5 | 84.3 |
| TIR-Bench 2/7 models | vision_language | — | — | — | 59.8 | — | 55.5 | — |
| Global PIQA 2/7 models | multilingual | — | — | — | 87.5 | — | 86.6 | — |
| AgentWorldBench Search 1/7 models | general_agent | — | — | — | — | — | 25.98 | — |
| IFEval 3/7 models | instruction_following | — | — | 91.5 | 95 | — | 91.9 | — |
| HLE (with tools) 5/7 models | stem_reasoning | 28.9 | — | — | 48.5 | 35.9 | 47.4 | 30.8 |
| LVBench 5/7 models | video_understanding | 71.4 | — | — | 73.6 | 76.6 | 71.4 | 72.4 |
| CharXiv (RQ) 6/7 models | document_understanding | 78 | 78.4 | — | 79.5 | 84.6 | 77.5 | 83.7 |
| GDP.pdf 2/7 models | long_context | 6.6 | — | — | — | — | — | 11.2 |
| BrowseComp 4/7 models | general_agent | 62 | — | 8.3 | 61 | — | 61 | — |
| MedXpertQA-MM 2/7 models | vision_language | — | — | — | 62.4 | — | 61.4 | — |
| MMLU-Redux 4/7 models | knowledge | 93.3 | 93.5 | — | 93.2 | — | 93.3 | — |
| NL2Repo-Bench 2/7 models | coding_agent | — | — | — | — | 48.1 | — | 42.3 |
| SkillsBench 1/7 models | general_agent | — | 46.6 | — | — | — | — | — |
| WebArena-Verified 2/7 models | general_agent | — | 48.8 | — | — | — | — | 64.8 |
| ParseBench Text Content 1/7 models | document_understanding | 90.7 | — | — | — | — | — | — |
| CoWorkBench 3/7 models | general_agent | — | 61 | — | — | 73.9 | — | 70.7 |
| Lab Bench (ProtocolQA) 1/7 models | safety | — | 69.1 | — | — | — | — | — |
| MCPMark 4/7 models | general_agent | 37 | — | 13.4 | 36.3 | — | 27 | — |
| AA-Omniscience Index (public set) 2/7 models | knowledge | -7.2 | — | — | — | — | — | -0.5 |
| MMVU 2/7 models | video_understanding | — | — | — | 73.3 | — | 72.3 | — |
| FullStackBench zh 2/7 models | coding_agent | — | — | — | 57.4 | — | 55 | — |
| MLVU 4/7 models | video_understanding | 86.2 | 86.6 | — | 85.9 | — | 85.6 | — |
| Beam128K 1/7 models | long_context | — | 63 | — | — | — | — | — |
| FullStackBench en 2/7 models | coding_agent | — | — | — | 60.1 | — | 58.1 | — |
| SWE-bench Verified 6/7 models | coding_agent | 73.4 | 77.2 | 53.1 | 75 | — | 70 | 74.6 |
| IMOAnswerBench 5/7 models | stem_reasoning | 78.9 | 80.8 | 69.8 | 79.9 | — | 76.8 | — |
| Humanity's Last Exam 6/7 models | stem_reasoning | 21.4 | 24 | 14.3 | 24.3 | — | 22.4 | 30.8 |
| CountBench 4/7 models | spatial_intelligence | 96.1 | 97.8 | — | 97.8 | — | 97.8 | — |
| WMDP (Chem) 1/7 models | safety | — | 74.8 | — | — | — | — | — |
| ERQA 6/7 models | spatial_intelligence | 61.8 | 62.5 | — | 60.5 | 72.3 | 64.8 | 65.5 |
| VideoMME (w/o sub.) 3/7 models | video_understanding | 82.5 | — | — | 82.8 | — | 82.5 | — |
| Terminal-Bench 2.0 4/7 models | coding_agent | 51.5 | 59.3 | — | 41.6 | — | 40.5 | — |
| DeepPlanning 3/7 models | general_agent | 25.9 | — | — | 22.6 | — | 22.8 | — |
| QwenSWEBench 2/7 models | coding_agent | — | 49.3 | — | — | — | — | 79 |
| QwenWebBench 4/7 models | coding_agent | 1397 | 1487 | — | 1068 | — | 978 | — |
| TAU2-Bench 3/7 models | general_agent | — | — | 79.1 | 79 | — | 81.2 | — |
| ODInW13 3/7 models | spatial_intelligence | 50.8 | — | — | 41.1 | — | 42.6 | — |
| VlmsAreBlind 4/7 models | vision_language | 96.6 | 97 | — | 96.9 | — | 97 | — |
| Terminal-Bench 2.1 (Terminus-2) 5/7 models | coding_agent | 52.5 | 63.4 | 21.3 | — | — | 41.4 | 73 |
| DeepSWE 1/7 models | coding_agent | 0 | — | — | — | — | — | — |
| OmniDocBench 1.5 5/7 models | document_understanding | 89.9 | 89.4 | — | 88.9 | — | 89.3 | 91.1 |
| HallusionBench 3/7 models | vision_language | 69.8 | — | — | 70 | — | 67.9 | — |
| OCRBench 4/7 models | document_understanding | 90 | 89.4 | — | 89.4 | — | 89.4 | — |
| HMMT Feb 26 6/7 models | stem_reasoning | 83.6 | 84.3 | 70.8 | 84.3 | — | 78.7 | 95.8 |
| Terminal-Bench Lite 2/7 models | coding_agent | 63.3 | — | — | — | — | — | 67 |
| AgentWorldBench OS 1/7 models | general_agent | — | — | — | — | — | 56.27 | — |
| AgentWorldBench MCP 1/7 models | general_agent | — | — | — | — | — | 57.87 | — |
| GPQA 1/7 models | reasoning | — | — | 77.2 | — | — | — | — |
| AgentWorldBench SWE 1/7 models | general_agent | — | — | — | — | — | 47.58 | — |
| CC-OCR 4/7 models | document_understanding | 81.9 | 81.2 | — | 81 | — | 80.7 | — |
| RefCOCO (avg) 4/7 models | spatial_intelligence | 92 | 92.5 | — | 90.9 | — | 89.2 | — |
| JobBench 3/7 models | general_agent | — | 21.8 | — | — | 55.7 | — | 33.4 |
| Artificial Analysis Intelligence Index 1/7 models | composite | 32 | — | — | — | — | — | — |
| SciCode (subtask) 1/7 models | stem_reasoning | — | 39.8 | — | — | — | — | — |
| Tool Decathlon 3/7 models | general_agent | 26.9 | — | — | 31.5 | — | 28.7 | — |
| Frontier-Bench v0.1 1/7 models | coding_agent | 1.4 | — | — | — | — | — | — |
| OSWorld 2.0 (Partial) 2/7 models | general_agent | — | — | — | — | 52.3 | — | 48 |
| ParseBench Mean 1/7 models | document_understanding | 44.1 | — | — | — | — | — | — |
| OSWorld 2.0 (Binary) 2/7 models | general_agent | — | — | — | — | 19.4 | — | 19.4 |
| ZEROBench_sub 3/7 models | vision_language | 34.4 | — | — | 36.2 | — | 34.1 | — |
| C-Eval 4/7 models | knowledge | 90 | 91.4 | — | 90.5 | — | 90.2 | — |
| WideSearch 3/7 models | general_agent | 60.1 | — | — | 66.4 | — | 59.1 | — |
| SimpleVQA 4/7 models | vision_language | 58.9 | 56.1 | — | 56 | — | 58.3 | — |
| VideoMME (w sub.) 4/7 models | video_understanding | 86.6 | 87.7 | — | 87 | — | 86.6 | — |
| Workspace Bench 1/7 models | — | — | — | 25.5 | — | — | — | — |
| SkillsBench Avg5 4/7 models | coding_agent | 28.7 | 48.2 | — | 27.2 | — | 4.4 | — |
| Vision2Web 3/7 models | vision_language | — | 45 | — | — | 64 | — | 62.9 |
| TauBench V3 Banking 1/7 models | general_agent | — | 16.7 | — | — | — | — | — |
| Claw-Eval Pass^3 6/7 models | coding_agent | 50 | 60.6 | — | 46.2 | 64.4 | 51 | 57.4 |
| WMT24++ (en→xx) 2/7 models | multilingual | — | — | — | 77.6 | — | 76.3 | — |
| LongBench v2 2/7 models | long_context | — | — | — | 60.6 | — | 59 | — |
| MCP-Atlas 5/7 models | general_agent | 62.8 | 62.5 | 47.4 | 68.4 | — | 62.4 | — |
| SWE Atlas - RF 3/7 models | coding_agent | 11.4 | — | 4.3 | — | — | 10.2 | — |
| V-Star 4/7 models | vision_language | 90.1 | 94.7 | — | 93.7 | — | 92.7 | — |
| AgentWorldBench Android 1/7 models | general_agent | — | — | — | — | — | 53.18 | — |
| DeepSWE 1.1 3/7 models | coding_agent | — | 13.3 | — | — | 58.7 | — | 42.2 |
| Terminal-Bench 2.1 (Claude Code) 3/7 models | coding_agent | 49.2 | — | 18.9 | — | — | 38.9 | — |
| Chatbot Korean (internal) 2/7 models | multilingual | 74.5 | — | — | — | — | — | 90.1 |
| MBCT 1/7 models | safety | — | 45.9 | — | — | — | — | — |
| VCT 1/7 models | safety | — | 33.7 | — | — | — | — | — |
| Gaokao 2026 1/7 models | — | — | — | 135.5 | — | — | — | — |
| Gaia2 1/7 models | general_agent | — | 40 | — | — | — | — | — |
| SUNRGBD 2/7 models | spatial_intelligence | — | — | — | 35.4 | — | 33.4 | — |
| SciCode 3/7 models | — | 35.6 | — | 32.7 | — | — | — | 44.6 |
| RefSpatialBench 4/7 models | spatial_intelligence | 64.3 | 70 | — | 67.7 | — | 63.5 | — |
| BrowseComp Plus 2/7 models | general_agent | 34.8 | — | — | — | — | — | 58 |
| MMMU 4/7 models | vision_language | 81.7 | 82.9 | — | 82.3 | — | 81.4 | — |
| GDPVal-AA v2 1/7 models | general_agent | — | 1141 | — | — | — | — | — |
| MVBench 4/7 models | video_understanding | 74.6 | 75.5 | — | 74.6 | — | 74.8 | — |
| ZEROBench 2/7 models | vision_language | — | — | — | 10 | — | 8 | — |
Comparison FAQ
Which model scores highest in this comparison?
Qwen3.8-Flash-Next wins this comparison with an overall score of 58 vs 52 for Qwen3.5-27B, leading 3 of 4 shared benchmarks.
Which model is strongest at General Agents?
Qwen3.8-Flash-Next leads the General Agents category (strength score 70/100 across 27 comparable benchmarks).
How reliable is this comparison verdict?
Based on 4 of 157 benchmarks scored by all 7 selected models.