Model comparison
GLM-4.5-Air vs DeepSeek-R1-Distill-Qwen-14B
- Input
- $0.20
- Output
- $1.10
- Context
- 128K
- Released
- 2025-07-28
- Input
- $0.18 / 1M tokens (Together AI)
- Output
- $0.18 / 1M tokens (Together AI)
- Context
- 128K (some serving platforms list 131K / 131,072 tokens)
- Released
- 2025-01-20
GLM-4.5-Air scores higher on 4 of the 4 benchmarks both models report; DeepSeek-R1-Distill-Qwen-14B is cheaper on input tokens ($0.2 vs $0.2 per 1M, 1.1×).
Specs and pricing side by side
| Spec | GLM-4.5-Air | DeepSeek-R1-Distill-Qwen-14B |
|---|---|---|
| Maker | Z.ai (Zhipu / GLM) | DeepSeek |
| Released | 2025-07-28 | 2025-01-20 |
| Context window | 128K | 128K (some serving platforms list 131K / 131,072 tokens) |
| Max output | 96K | 32,768 tokens (DeepSeek's recommended maximum… |
| Input price | $0.20 | $0.18 / 1M tokens (Together AI) |
| Output price | $1.10 | $0.18 / 1M tokens (Together AI) |
| Parameters | 106B total · 12B active | 14B (dense; based on Qwen2.5-14B) |
| License | MIT | MIT License (DeepSeek code repository and model weights). The model is derived from Qwen2.5-14B, which is originally released under the Apache 2.0 License; commercial use and modification are permitted. |
| Weights | Open weights | Open weightssame |
| Modalities | Text | Textsame |
| Knowledge cutoff | Not disclosed | Not separately disclosed by DeepSeek; the… |
Benchmarks both models report
| Benchmark | GLM-4.5-Air | DeepSeek-R1-Distill-Qwen-14B |
|---|---|---|
| AIME 2024 | 89.4% | 69.7% |
| MATH-500 | 98.1% | 93.9% |
| GPQA Diamond | 71.72% | 59.1% |
| LiveCodeBench | 70.7% | 53.1% |
Scores are as published by each maker or its cited source; harness, reasoning effort and date can differ between rows, so treat small gaps as ties. Sources are listed on each model page.
Which one should you pick?
Pick GLM-4.5-Air if…
- You care about AIME 2024, MATH-500, GPQA Diamond (it scores higher on 4 shared benchmarks).
- Your use case is: agentic coding assistants and autonomous coding agents that call tools over long horizons.
Pick DeepSeek-R1-Distill-Qwen-14B if…
- Cost matters: input tokens cost 1.1× less ($0.2 vs $0.2 per 1M).
- Your use case is: self-hosted, privacy-sensitive reasoning workloads where a small open-weight model with visible chain-of-thought is preferred over a closed API.
FAQ
Which is better, GLM-4.5-Air or DeepSeek-R1-Distill-Qwen-14B?
GLM-4.5-Air scores higher on 4 of the 4 benchmarks both models report; DeepSeek-R1-Distill-Qwen-14B is cheaper on input tokens ($0.2 vs $0.2 per 1M, 1.1×). The better pick depends on whether you value benchmark scores, price per token, context length or openness for your workload.
How much cheaper is DeepSeek-R1-Distill-Qwen-14B?
GLM-4.5-Air: $0.20 input, $1.10 output. DeepSeek-R1-Distill-Qwen-14B: $0.18 / 1M tokens (Together AI) input, $0.18 / 1M tokens (Together AI) output (list API prices as published by each maker).
What are the context windows of GLM-4.5-Air and DeepSeek-R1-Distill-Qwen-14B?
GLM-4.5-Air offers 128K with up to 96K output; DeepSeek-R1-Distill-Qwen-14B offers 128K (some serving platforms list 131K / 131,072 tokens) with up to 32,768 tokens (DeepSeek's recommended maximum generation length for R1 models) output.
Is GLM-4.5-Air newer than DeepSeek-R1-Distill-Qwen-14B?
GLM-4.5-Air was released 2025-07-28; DeepSeek-R1-Distill-Qwen-14B was released 2025-01-20.