Decision Workspace
ferrite-inference vs bebelm vs rage-quant
Side-by-side comparison of Rust crates
48
ferrite-inference
experimentalv0.2.0
CPU-native GGUF inference kernels and runtime for Ferrite
48
bebelm
experimentalv0.3.6
CPU-only, pure-Rust inference for Liquid AI LFM2.5-8B-A1B (Q4_K_M)
42
rage-quant
experimentalv0.1.0
High-performance quantized GEMV kernels for CPU-only LLM inference. Direct dot product on Q8_0/Q6_K/Q4_K GGUF blocks with AVX2+FMA SIMD — 3.0x decode speedup.
Core Metrics
| ferrite-inference | bebelm | rage-quant | |
|---|---|---|---|
| Health Score | 48 | 48 | 42 |
| Total Downloads | 32 | 167 | 23 |
| 30d Downloads | 0 | 0 | 0 |
| Dependents | 0 | 0 | 0 |
| Releases | 2 | 8 | 1 |
| Last Updated | today | 31d ago | 88d ago |
| Age | 3d | 1m | 2m |
Health Breakdown
ferrite-inference
Maintenance
14
Quality
14
Community
6
Popularity
2
Documentation
12
bebelm
Maintenance
17
Quality
12
Community
6
Popularity
3
Documentation
10
rage-quant
Maintenance
8
Quality
13
Community
6
Popularity
2
Documentation
13
Technical Details
| ferrite-inference | bebelm | rage-quant | |
|---|---|---|---|
| Version | 0.2.0 | 0.3.6 | 0.1.0 |
| Stable (≥1.0) | ✗ No | ✗ No | ✗ No |
| License | Apache-2.0 | MIT | AGPL-3.0-only |
| Dependencies | 5 | 4 | 5 |
| Crate Size | 97KB | 112KB | 30KB |
| Features | 1 | 0 | 0 |
| Yanked % | 0.0% | 0.0% | 0.0% |
| Edition | 2021 | 2021 | 2021 |
| MSRV | 1.96 | — | — |
| Owners | 1 | 1 | 1 |
Links
Quick Verdict
- •ferrite-inference leads with a health score of 48/100, but none of the options score above 80.
- •bebelm has the most downloads (167), suggesting wider adoption.