Real data comparison
Tencent: Hy4 preview vs MiniMax: MiniMax M3 (batch)
On price, MiniMax: MiniMax M3 (batch) is cheaper ($0.52 vs $1.25 per 1M). In real production usage, Tencent: Hy4 preview is used more (rank #3 vs #9 of 325).
- Tencent: Hy4 preview
- MiniMax: MiniMax M3 (batch)
Head-to-head by real task
Tencent: Hy4 preview leads 6 of 13 shared task categories by real classified traffic share; MiniMax: MiniMax M3 (batch) leads 6. Not a benchmark score - this is what developers actually route to each model for, per OpenRouter's task classification.
| Task | Tencent: Hy4 preview | MiniMax: MiniMax M3 (batch) |
|---|---|---|
| File I/O | 20.2% | 3.7% |
| Workflow Execution | 14.1% | 2.6% |
| Multi-step Planning | 6.2% | 2.8% |
| Web Search | 3.5% | 4.8% |
| Debugging | 2.7% | 3.8% |
| Code Generation | 3.1% | 2.2% |
| Finance & Trading | 2.8% | 3.7% |
| Frontend & UI | 4.5% | 5.2% |
| DevOps & Config | 2.1% | 2.7% |
| Repo Scanning | 4.0% | 3.7% |
| SQL & Database | 3.2% | 3.1% |
| Memory Extraction | 2.4% | 2.5% |
| Shell Execution | 3.1% | 3.1% |
Usage share over time
- Tencent: Hy4 preview
- MiniMax: MiniMax M3 (batch)
Frequently asked questions
Tencent: Hy4 preview vs MiniMax: MiniMax M3 (batch): which is better quality?
Quality data is unavailable for one or both models.
Tencent: Hy4 preview vs MiniMax: MiniMax M3 (batch): which is cheaper?
MiniMax: MiniMax M3 (batch) is cheaper at $0.52/1M vs $1.25/1M.
Tencent: Hy4 preview vs MiniMax: MiniMax M3 (batch): which is used more?
Tencent: Hy4 preview has higher real OpenRouter usage share (rank #3 vs #9 of 325).
Tencent: Hy4 preview vs MiniMax: MiniMax M3 (batch): which wins more real-world task categories?
Tencent: Hy4 preview and MiniMax: MiniMax M3 (batch) split the 13 shared task categories evenly.
Source: OpenRouter (openrouter.ai/rankings), as of September 9, 2026.
Benchmark scores: Artificial Analysis (artificialanalysis.ai) via OpenRouter (openrouter.ai/rankings).
Methodology: www.smophy.ai/benchmark/methodology
Token counts originate from each provider's own tokenizer and are not directly comparable across providers.