Real data comparison
Tencent: Hy3 vs Xiaomi: MiMo-V2.5
On price, Xiaomi: MiMo-V2.5 is cheaper ($0.18 vs $0.23 per 1M). In real production usage, Xiaomi: MiMo-V2.5 is used more (rank #2 vs #3 of 291). Across 16 real task categories where both models have classified traffic, Xiaomi: MiMo-V2.5 handles more of the traffic in 10 of them.
- Tencent: Hy3
- Xiaomi: MiMo-V2.5
Head-to-head by real task
Tencent: Hy3 leads 6 of 16 shared task categories by real classified traffic share; Xiaomi: MiMo-V2.5 leads 10. Not a benchmark score - this is what developers actually route to each model for, per OpenRouter's task classification.
| Task | Tencent: Hy3 | Xiaomi: MiMo-V2.5 |
|---|---|---|
| Repo Scanning | 4.8% | 18.1% |
| Frontend & UI | 7.4% | 14.7% |
| DevOps & Config | 6.3% | 12.0% |
| Workflow Execution | 10.4% | 4.8% |
| SQL & Database | 3.1% | 8.3% |
| Debugging | 4.2% | 9.0% |
| Research & Reports | 6.7% | 2.3% |
| Memory Extraction | 5.0% | 8.7% |
| Multi-step Planning | 6.7% | 3.4% |
| File I/O | 13.7% | 10.7% |
| Code Generation | 5.3% | 8.1% |
| DevOps | 8.9% | 11.3% |
| Web Search | 5.2% | 3.0% |
| Shell Execution | 6.0% | 5.2% |
| Code Review | 3.6% | 4.1% |
| Conversation | 4.1% | 4.2% |
Usage share over time
- Tencent: Hy3
- Xiaomi: MiMo-V2.5
Frequently asked questions
Tencent: Hy3 vs Xiaomi: MiMo-V2.5: which is better quality?
Quality data is unavailable for one or both models.
Tencent: Hy3 vs Xiaomi: MiMo-V2.5: which is cheaper?
Xiaomi: MiMo-V2.5 is cheaper at $0.18/1M vs $0.23/1M.
Tencent: Hy3 vs Xiaomi: MiMo-V2.5: which is used more?
Xiaomi: MiMo-V2.5 has higher real OpenRouter usage share (rank #2 vs #3 of 291).
Tencent: Hy3 vs Xiaomi: MiMo-V2.5: which wins more real-world task categories?
Xiaomi: MiMo-V2.5 handles more traffic in 10 of 16 shared task categories where both models have classified OpenRouter usage.
Source: OpenRouter (openrouter.ai/rankings), as of August 9, 2026.
Benchmark scores: Artificial Analysis (artificialanalysis.ai) via OpenRouter (openrouter.ai/rankings).
Methodology: smophy.ai/benchmark/methodology
Token counts originate from each provider's own tokenizer and are not directly comparable across providers.