ZwZ2-27B on local hardware
Estimated comfort for repository coding with Aider on a MacBook Pro M4 Pro 48GB.
ZwZ2-27B
inclusionAI · See model card · 121 downloads
VisionReasoning / thinking mode
- Parameters
- 27B exact
- Context window
- 256K tokens
- Attention
- Hybrid linear + full attention
- Layers · KV heads · head dim
- 64 · 4 · 256
- KV cache
- 64 KB/token · 2.0 GB at 32K
- Tool calling
- good (from chat template)
- Thinking
- good (from chat template)
- Vision
- good (from config)
- Commercial use
- check cardCheck the model card
- Import confidence
- medium · capabilities estimated
Weights by quantization
Q4_K_M
15 GB
gguf file size
Q5_K_M
18 GB
gguf file size
Q6_K
21 GB
gguf file size
F16 / BF16
50 GB
gguf file size
- • vision: good, inferred from config (low confidence).
- • Capability ratings are estimated from model size and name — this model hasn't been assessed by hand.
- • Commercial use unknown: Check the model card.
- • GGUF sizes come from community conversions matched by model name; equivalence and runtime support are not verified. MLX conversion availability is unverified.
ZwZ2-27B — 4-bit
Borderline for your workload
Speed
10–14 tok/s
Slow
Memory left
19 GB
of 48 GB
First word in
≈2 min
Too slow
All the details
- Your setup
- MacBook Pro M4 Pro 48 GB
- Workload
- Repository coding with Aider
- Runtime → tool
- LM Studio → Aider
- Memory used
- 18 GB
- Context
- 32K · up to 64K practical
- Tool support
- Official
- Use-case fit
- Partly suited
Why?
- Model fits with approximately 19 GB of memory headroom.
- LM Studio supports hardware acceleration on this machine (Metal (Apple GPU)).
- Aider can connect to LM Studio (officially supported).
- 32K window; generation drops to 10–13 tok/s when it is full.
- Coding capability is below what interactive coding ideally needs — expect more mistakes or retries.
- Generation (10–14 tok/s) is slow for interactive coding.
- Ingesting a prompt of 14K tokens takes ≈2 min and each step ≈2 min; across ~4 calls a task takes ≈10 min.
Why “Borderline” rather than “Acceptable”?
- Prompt processing: Ingesting the first prompt (14K tokens) takes ≈2 min; each of the ~4 follow-up turns takes ≈2 min (≈10 min per task).
- Generation: Expected 10–14 tok/s at ~16K tokens of context. Coding with repository context is comfortable from about 28 tok/s.
- Workload suitability: Coding capability is below what interactive coding ideally needs — expect more mistakes or retries.
- Context: 32K window; generation drops to 10–13 tok/s when it is full.
How sure are we?
Medium confidenceEstimated from bandwidth and compute, calibrated against 2 verified benchmarks of other models on this same chip (measured with a different engine, so calibration is partial).
- Calibrated with verified benchmarks of other models on this chip.
- Hybrid/sliding attention performance varies more between runtimes.
Medium confidence