GLM-4.7-Flash on local hardware
Estimated comfort for repository coding with Aider on a MacBook Pro M4 Pro 48GB.
GLM-4.7-FlashOpen source
zai-org · mit · 1,125,508 downloads
Mixture of expertsReasoning / thinking mode
- Parameters
- 31B exact
- Active per token
- ≈4.9B (4 of 64 experts)
- Context window
- 198K tokens
- Attention
- Multi-head latent attention (compressed KV)
- Layers · KV heads · head dim
- 47 · 20 · 102
- KV cache
- 53 KB/token · 1.7 GB at 32K
- Tool calling
- good (from chat template)
- Thinking
- good (from chat template)
- Vision
- basic (inferred — no supporting signal)
- Commercial use
- allowedRetain the required license and attribution notices
- Import confidence
- medium · capabilities estimated
Weights by quantization
Q3_K_M
14 GB
gguf file size
Q4_K_M
17 GB
gguf file size
Q5_K_M
20 GB
gguf file size
Q6_K
23 GB
gguf file size
Q8_0
30 GB
gguf file size
F16 / BF16
56 GB
gguf file size
MXFP4
16 GB
gguf file size
- • vision: basic, inferred from missing signals (low confidence).
- • Capability ratings are estimated from model size and name — this model hasn't been assessed by hand.
- • GGUF sizes come from community conversions matched by model name; equivalence and runtime support are not verified. MLX conversion availability is unverified.
GLM-4.7-Flash — MXFP4
Acceptable for your workload
Speed
35–50 tok/s
Excellent
Memory left
18 GB
of 48 GB
First word in
≈30 s
Comfortable
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 128K practical
- Tool support
- Official
- Use-case fit
- Partly suited
Why?
- Model fits with approximately 18 GB of memory headroom.
- LM Studio supports hardware acceleration on this machine (Metal (Apple GPU)).
- Aider can connect to LM Studio (officially supported).
- Generation (35–50 tok/s) should be appropriate for interactive coding.
- Ingesting a prompt of 14K tokens takes ≈30 s; prompt processing keeps up with interactive coding.
- 32K working context is reasonable.
- Weak coding capability for interactive coding.
Why “Acceptable” rather than “Comfortable”?
- Workload suitability: Weak coding capability for interactive coding.
- Prompt processing: Ingesting the first prompt (14K tokens) takes ≈30 s; each of the ~4 follow-up turns takes ≈30 s (≈3 min per task).
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.
- MoE performance varies more between runtimes.
Medium confidence