Decision Workspace
zeta-quantization vs bitnet-quantize vs tq-kv
Side-by-side comparison of Rust crates
45
zeta-quantization
experimentalv0.1.0
Advanced quantization engine for efficient LLM inference
52
bitnet-quantize
experimentalv0.2.1
Microsoft BitNet b1.58 quantization and inference for Rust
52
tq-kv
experimentalv0.4.0
TurboQuant: Extreme KV Cache Compression for LLMs — ICLR 2026 Pure Rust implementation. Lloyd-Max codebook, fused attention, SIMD (AVX2), candle KvCache drop-in. 2-4 bit, up to 15x compression. Supports Qwen2, Llama, Mistral, Phi-3, Gemma2.
Core Metrics
| zeta-quantization | bitnet-quantize | tq-kv | |
|---|---|---|---|
| Health Score | 45 | 52 | 52 |
| Total Downloads | 651 | 528 | 44 |
| 30d Downloads | 52 | 164 | 44 |
| Dependents | 3 | 12 | 0 |
| Releases | 1 | 3 | 4 |
| Last Updated | 186d ago | 57d ago | today |
| Age | 6m | 2m | 1d |
Health Breakdown
zeta-quantization
Maintenance
5
Quality
14
Community
8
Popularity
3
Documentation
15
bitnet-quantize
Maintenance
11
Quality
14
Community
9
Popularity
3
Documentation
15
tq-kv
Maintenance
19
Quality
15
Community
6
Popularity
2
Documentation
10
Technical Details
| zeta-quantization | bitnet-quantize | tq-kv | |
|---|---|---|---|
| Version | 0.1.0 | 0.2.1 | 0.4.0 |
| Stable (≥1.0) | ✗ No | ✗ No | ✗ No |
| License | MIT OR Apache-2.0 | MIT | MIT OR Apache-2.0 |
| Dependencies | 6 | 12 | 5 |
| Crate Size | 18KB | 58KB | 45KB |
| Features | 1 | 4 | 4 |
| Yanked % | 0.0% | 0.0% | 0.0% |
| Edition | 2021 | 2021 | 2021 |
| MSRV | 1.70 | 1.92 | — |
| Owners | 1 | 1 | 1 |
Links
Quick Verdict
- •bitnet-quantize leads with a health score of 52/100, but none of the options score above 80.
- •zeta-quantization has the most downloads (651), suggesting wider adoption.