GLM-4.7-Flash vs Qwen3-Coder-30B-A3B-Instruct
GLM-4.7-Flash (31.2B parameters) and Qwen3-Coder-30B-A3B-Instruct (30.5B parameters) side by side: the memory each needs at every precision, what it costs to run on a live GPU, and the context window, KV cache and license where they are published. Numbers are computed from the models' published specs; this page does not rank quality.
Side by side
| Fact | GLM-4.7-Flash | Qwen3-Coder-30B-A3B-Instruct |
|---|---|---|
| Parameters | 31.2B (Mixture-of-experts: 4 of 64 experts active per token (exact active-parameter count not stated on the model card)) | 30.5B |
| Architecture | Multi-head latent attention; mixture of 64 experts, 4 active per token | Grouped-query attention; mixture of 128 experts, 8 active per token |
| Context length | 198K tokens (202,752) | 262,144 tokens |
| License | MIT | – |
| Published precision | BF16 | BF16 |
| VRAM needed, As published | 69.8 GB | 68.2 GB |
| VRAM needed, FP8 | 34.9 GB | 34.1 GB |
| VRAM needed, INT4 | 17.4 GB | 17.1 GB |
| Cheapest live fit, As published | A100 · $1.21/hr | A100 · $1.21/hr |
| Cheapest live fit, FP8 | L40 · $0.742/hr | L40 · $0.742/hr |
| Cheapest live fit, INT4 | RTX A5000 · $0.176/hr | RTX A5000 · $0.176/hr |
| KV cache per token (16-bit) | 53 KB | 96 KB |
| KV cache at 32k tokens | 1.65 GB | 3.00 GB |
| KV cache at 128k tokens | 6.61 GB | 12.0 GB |
VRAM is the weight size at each precision times a flat 1.2 overhead; see the methodology. The FP8 and INT4 rows need a quantized checkpoint or an engine that quantizes on load. The fit is the lowest-priced single GPU type that holds the model at that precision, or the lowest-priced multi-GPU set (up to 8) when none does. KV cache is for one sequence at 16-bit, computed from each model's config where the attention layout is known.
Which to pick
- Qwen3-Coder-30B-A3B-Instruct needs less VRAM at its published precision (68.2 GB against 69.8 GB), so it fits on a smaller GPU.
- Qwen3-Coder-30B-A3B-Instruct lists the longer context window (262,144 tokens against 202,752).
- GLM-4.7-Flash caches less per sequence at 32k tokens (1.7 GB against 3.0 GB), leaving more memory for batching.
These follow only from the facts in the table above. Whether either model does your task well is a separate question this page does not answer.
Keep reading
- GLM-4.7-Flash: full VRAM table and live GPU fit
- Qwen3-Coder-30B-A3B-Instruct: full VRAM table and live GPU fit
- The GLM model series
- The Qwen model series
- All models that fit in 80 GB
Other comparisons
