Qwen2.5 VL 72B Instruct is 2.5× cheaper per million tokens (blended). Gemini 3.7 Flash has the larger context window (1,048,576 vs 128,000 tokens, 8.2×). Qwen2.5 VL 72B Instruct is open-weights; the other is not.
Gemini 3.7 Flash vs Qwen2.5 VL 72B Instruct
Live catalog fields. Best-in-row is highlighted.
Gemini 3.7 Flash
Google · google__gemini-3.7-flash
Qwen2.5 VL 72B Instructopen
Qwen · qwen__qwen2.5-vl-72b-instruct
Identity
| Field | google__gemini-3.7-flash | qwen__qwen2.5-vl-72b-instruct |
|---|---|---|
| Provider | Qwen | |
| Slug | google__gemini-3.7-flash | qwen__qwen2.5-vl-72b-instruct |
| Status | live | live |
| Open weights | No | Yes |
| License | — | — |
| HuggingFace | — | Qwen/Qwen2.5-VL-72B-Instruct |
Quality
| Field | google__gemini-3.7-flash | qwen__qwen2.5-vl-72b-instruct |
|---|---|---|
| Quality index (0–100) | 86.7 | — |
Cost
| Field | google__gemini-3.7-flash | qwen__qwen2.5-vl-72b-instruct |
|---|---|---|
| Input $ / 1M tokens | $0.75 | $0.80 |
| Cached input $ / 1M | $0.07 | $0.40 |
| Output $ / 1M tokens | $3.75 | $1.00 |
Context
| Field | google__gemini-3.7-flash | qwen__qwen2.5-vl-72b-instruct |
|---|---|---|
| Max context tokens | 1,048,576 | 128,000 |
| Max output tokens | 65,536 | 115,200 |
Modalities
| Field | google__gemini-3.7-flash | qwen__qwen2.5-vl-72b-instruct |
|---|---|---|
| Input | text, image, video, file, audio | text, image |
| Output | text | text |