Qwen3.8-Flash-Next vs Edge0-145B-NVFP4
New ComparisonScores are fetched live from the current database on every visit — nothing is cached.
Winner
Qwen3.8-Flash-Next wins this comparison with an overall score of 84 vs 69 for Edge0-145B-NVFP4, leading 3 of 6 shared benchmarks.
Based on 6 of 27 benchmarks scored by all 2 selected models.
Benchmark Scores Comparison
Qwen3.8-Flash-Next
125.0B params
Score: 70.0
Edge0-145B-NVFP4
145.0B params
Coverage matrix
2 models × 27 benchmarks — hover a benchmark name for its category
| Benchmark | Qwen3.8-Flash-Next Qwen Team (Alibaba) | Edge0-145B-NVFP4 Edge0 |
|---|---|---|
| GPQA Diamond | 91.7 | 90.4 |
| LiveCodeBench v6 | 91.9 | — |
| SWE-bench Pro | 62.5 | — |
| RealWorldQA | 88.5 | — |
| Agents' Last Exam (Score) | 51.2 | — |
| AA-LCR | 79.1 | 79.3 |
| MathVision | 90.6 | — |
| AndroidWorld | 84.5 | — |
| Agents' Last Exam | 24.3 | — |
| Toolathlon Verified | 73.5 | — |
| RecreationBench | 49.9 | — |
| Claw-Eval Avg | 60.4 | — |
| HLE (with tools) | 35.9 | 35.4 |
| LVBench | 76.6 | — |
| CharXiv (RQ) | 84.6 | — |
| NL2Repo-Bench | 48.1 | — |
| CoWorkBench | 73.9 | — |
| SWE-bench Multilingual | 81 | — |
| IFBench | 81.3 | 80.6 |
| ERQA | 72.3 | — |
| Terminal-Bench 2.1 (Terminus-2) | 83.3 | 86.5 |
| JobBench | 55.7 | — |
| OSWorld 2.0 (Partial) | 52.3 | — |
| OSWorld 2.0 (Binary) | 19.4 | — |
| Vision2Web | 64 | — |
| Claw-Eval Pass^3 | 64.4 | — |
| DeepSWE 1.1 | 58.7 | 64.6 |
| Covered | 27/27 | 6/27 |
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 | 83.8 | 3 of 6 | 27 of 27 (100.0%) | STEM & Reasoning | Agentic Coding |
| Edge0-145B-NVFP4 | 68.9 | 3 of 6 | 6 of 27 (22.0%) | Long Context | STEM & Reasoning |
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 | Edge0-145B-NVFP4 |
|---|---|---|
| Agentic Coding 7 benchmarks | 93 | 100 |
| STEM & Reasoning 3 benchmarks | 100 | 97.6 |
| Instruction Following 1 benchmark | 100 | 98.8 |
| Long Context 1 benchmark | 99.7 | 100 |
Fully comparable benchmarks
Scored by all 2 selected models (6)
| Benchmark | Category | Qwen3.8-Flash-Next | Edge0-145B-NVFP4 |
|---|---|---|---|
| GPQA Diamond | stem_reasoning | 91.7 | 90.4 |
| AA-LCR | long_context | 79.1 | 79.3 |
| HLE (with tools) | stem_reasoning | 35.9 | 35.4 |
| IFBench | instruction_following | 81.3 | 80.6 |
| Terminal-Bench 2.1 (Terminus-2) | coding_agent | 83.3 | 86.5 |
| DeepSWE 1.1 | coding_agent | 58.7 | 64.6 |
Partial coverage benchmarks
Scored by at least one but not all selected models (21) — missing scores show as —
| Benchmark | Category | Qwen3.8-Flash-Next | Edge0-145B-NVFP4 |
|---|---|---|---|
| LiveCodeBench v6 1/2 models | stem_reasoning | 91.9 | — |
| SWE-bench Pro 1/2 models | coding_agent | 62.5 | — |
| RealWorldQA 1/2 models | vision_language | 88.5 | — |
| Agents' Last Exam (Score) 1/2 models | general_agent | 51.2 | — |
| MathVision 1/2 models | vision_language | 90.6 | — |
| AndroidWorld 1/2 models | general_agent | 84.5 | — |
| Agents' Last Exam 1/2 models | general_agent | 24.3 | — |
| Toolathlon Verified 1/2 models | general_agent | 73.5 | — |
| RecreationBench 1/2 models | general_agent | 49.9 | — |
| Claw-Eval Avg 1/2 models | coding_agent | 60.4 | — |
| LVBench 1/2 models | video_understanding | 76.6 | — |
| CharXiv (RQ) 1/2 models | document_understanding | 84.6 | — |
| NL2Repo-Bench 1/2 models | coding_agent | 48.1 | — |
| CoWorkBench 1/2 models | general_agent | 73.9 | — |
| SWE-bench Multilingual 1/2 models | coding_agent | 81 | — |
| ERQA 1/2 models | spatial_intelligence | 72.3 | — |
| JobBench 1/2 models | general_agent | 55.7 | — |
| OSWorld 2.0 (Partial) 1/2 models | general_agent | 52.3 | — |
| OSWorld 2.0 (Binary) 1/2 models | general_agent | 19.4 | — |
| Vision2Web 1/2 models | vision_language | 64 | — |
| Claw-Eval Pass^3 1/2 models | coding_agent | 64.4 | — |
Comparison FAQ
Which model is better, Qwen3.8-Flash-Next or Edge0-145B-NVFP4?
Qwen3.8-Flash-Next wins this comparison with an overall score of 84 vs 69 for Edge0-145B-NVFP4, leading 3 of 6 shared benchmarks.
Which model is better at Agentic Coding, Qwen3.8-Flash-Next or Edge0-145B-NVFP4?
Edge0-145B-NVFP4 leads the Agentic Coding category (strength score 100/100 across 2 comparable benchmarks).
Where does Edge0-145B-NVFP4 outperform Qwen3.8-Flash-Next?
Edge0-145B-NVFP4 leads the Agentic Coding category (strength score 100/100 across 2 comparable benchmarks).
How reliable is this comparison verdict?
Based on 6 of 27 benchmarks scored by all 2 selected models.