Decision support
Wan 2.5 Preview vs MiniMax H3
Compare canonical objects directly. The useful dimensions change with the object type, so model and hardware comparisons stay separate.
Alibaba
Benchmark results
Head to head on 2 benchmarks both models report
- Wan 2.5 Preview0leads
- MiniMax H32leads
Arena Text to Video
Arena rating · Higher is better
- Wan 2.5 Preview1244Arena #27CI 1236 to 1253
- MiniMax H3Best1456+212 ahead · Arena #11CI 1448 to 1464
Runs and sources
Wan 2.5 Preview- default124495% CI 1236 to 1253 · 43,113 votes
MiniMax H3- default145695% CI 1448 to 1464 · 29,257 votes
Arena Image to Video
Arena rating · Higher is better
- Wan 2.5 Preview1322Arena #24CI 1312 to 1332
- MiniMax H3Best1493+171 ahead · Arena #3CI 1488 to 1499
Runs and sources
Wan 2.5 Preview- default132295% CI 1312 to 1332 · 17,048 votes
MiniMax H3- default149395% CI 1488 to 1499 · 121,481 votes
Arena Video Edit
Arena rating · Higher is better
- Wan 2.5 PreviewNot reported
- MiniMax H31378Arena #5CI 1369 to 1388
Runs and sources
Wan 2.5 PreviewNot reported
MiniMax H3- default137895% CI 1369 to 1388 · 8,385 votes
Each model's best published result is shown; "Not reported" means no matched evaluation, not a score of zero. Arena and DeepSWE ranks are the leaderboards' own; other ranks are among catalogue models with a score. Results from different harnesses or settings are not directly comparable.
Arena ratings: Arena (lmarena-ai/leaderboard-dataset), licensed CC BY 4.0. Ratings come from pairwise human votes and are only comparable within one leaderboard; the interval is Arena's 95% confidence interval. Model names are matched to the Vertical Tabs catalog.
Model specifications
API pricing · USD per 1M tokens
Capabilities and modalities
Architecture and availability
A dash means the field is not reported. Provider-specific performance and pricing may differ.