GLM 4.5 Air Derestricted vs GLM 4.7 Flash REAP 23B A3B
Side-by-side comparison of VRAM requirements, quantization, context length, and hardware compatibility.
Specifications
| GLM 4.5 Air Derestricted | GLM 4.7 Flash REAP 23B A3B | |
|---|---|---|
| Parameters | 110.5B | 23.0B |
| Context | 131K | 203K |
| Architecture | Glm4MoeForCausalLM | Glm4MoeLiteForCausalLM |
| License | MIT | MIT |
| Downloads | 172 | 542 |
| Released | Dec 2025 | Jan 2026 |
VRAM by Quantization: GLM 4.5 Air Derestricted vs GLM 4.7 Flash REAP 23B A3B
| Quantization | Bits | GLM 4.5 Air Derestricted VRAM | GLM 4.7 Flash REAP 23B A3B VRAM |
|---|---|---|---|
| Q2_K | 3.40 | 47.4 GB | 10.9 GB |
| Q3_K_M | 3.90 | 54.3 GB | 12.3 GB |
| Q3_K_S | 3.50 | 48.8 GB | 11.2 GB |
| Q4_0 | 4.00 | 55.7 GB | 12.6 GB |
| Q4_K_M | 4.80 | 66.7 GB | 14.9 GB |
| Q5_K_M | 5.70 | 79.1 GB | 17.5 GB |
| Q6_K | 6.60 | 91.6 GB | 20.1 GB |
| Q8_0 | 8.00 | 110.9 GB | 24.1 GB |
Verdict
GLM 4.7 Flash REAP 23B A3B needs less VRAM at Q4_K_M (14.9 GB vs 66.7 GB), so it fits on smaller GPUs. GLM 4.7 Flash REAP 23B A3B supports a longer context window (203K tokens). GLM 4.7 Flash REAP 23B A3B is the more widely downloaded of the two.
Frequently Asked Questions
- Which needs less VRAM, GLM 4.5 Air Derestricted or GLM 4.7 Flash REAP 23B A3B?
At Q4_K_M, GLM 4.5 Air Derestricted needs 66.7 GB and GLM 4.7 Flash REAP 23B A3B needs 14.9 GB, so GLM 4.7 Flash REAP 23B A3B is the lighter option to run locally.
- Which has a longer context window, GLM 4.5 Air Derestricted or GLM 4.7 Flash REAP 23B A3B?
GLM 4.5 Air Derestricted supports 131,072 tokens and GLM 4.7 Flash REAP 23B A3B supports 202,752 tokens.
- What is the difference between GLM 4.5 Air Derestricted and GLM 4.7 Flash REAP 23B A3B?
GLM 4.5 Air Derestricted is a 110.5B model from ArliAI (GLM family), while GLM 4.7 Flash REAP 23B A3B is a 23.0B model from Cerebras (GLM family). Compare their VRAM requirements above to see which fits your GPU or Mac.