# DeepSeek R1 0528 NVFP4 vs Qwen2.5-VL (72B) Instruct

On provider list prices, Qwen2.5-VL (72B) Instruct costs $1.95 per million input tokens against $3 for DeepSeek R1 0528 NVFP4: 1.5x apart. Output is $8 against $7. On Allocate both bill at list plus the 7% transaction fee.

## Specifications

| | DeepSeek R1 0528 NVFP4 | Qwen2.5-VL (72B) Instruct |
| --- | --- | --- |
| Lab | Deepseek | Qwen |
| Access | Open weights | Open weights |
| Context window | 160K tokens | 32K tokens |
| List price, input | $3 / M tokens | $1.95 / M tokens |
| List price, output | $7 / M tokens | $8 / M tokens |
| Cached input | n/a | n/a |
| License | MIT | Qwen license |
| Fine-tunable | Yes | Yes |

Specifications and provider list prices from the Allocate catalog, checked 2026-07-08. Billed price is list plus the 7% transaction fee.

## What the numbers say

Take 1,000,000 requests a month at 1,200 input and 350 output tokens each. That workload costs $5,140 a month on Qwen2.5-VL (72B) Instruct and $6,050 on DeepSeek R1 0528 NVFP4 at list: a gap of $910, or 1.2x.

DeepSeek R1 0528 NVFP4 reads 160K tokens per request against 32K for Qwen2.5-VL (72B) Instruct, 5.0x the window. That decides which one can take whole documents without splitting them.

## Choose DeepSeek R1 0528 NVFP4 for

- The longer context window (160K vs 32K tokens)
- Fine-tuning under a permissive license (MIT)

## Choose Qwen2.5-VL (72B) Instruct for

- The lower list price ($1.95 in / $8 out per M tokens)

## Common questions

### Which is cheaper, DeepSeek R1 0528 NVFP4 or Qwen2.5-VL (72B) Instruct?

Qwen2.5-VL (72B) Instruct, on this workload shape. At list prices it is $1.95/$8 per million tokens in and out against $3/$7 for DeepSeek R1 0528 NVFP4. Billed on Allocate: $2.09/$8.56 against $3.21/$7.49, list plus 7%.

### Which has the bigger context window?

DeepSeek R1 0528 NVFP4: 163,840 tokens (160K) against 32,768 (32K) for Qwen2.5-VL (72B) Instruct.

### Can I fine-tune DeepSeek R1 0528 NVFP4 or Qwen2.5-VL (72B) Instruct?

Both publish open weights (DeepSeek R1 0528 NVFP4: MIT; Qwen2.5-VL (72B) Instruct: Qwen license), so both can be fine-tuned. On Allocate the trained weights stay inside your boundary and belong to you.

---

[HTML page](https://allocate.network/compare/deepseek-deepseek-r1-0528-vs-qwen-qwen2-5-vl-72b-instruct) · [DeepSeek R1 0528 NVFP4](https://allocate.network/models/deepseek-deepseek-r1-0528.md) · [Qwen2.5-VL (72B) Instruct](https://allocate.network/models/qwen-qwen2-5-vl-72b-instruct.md) · [Machine-readable catalog](https://allocate.network/catalog.json)