AI Rate CardQwen 2.5 72B Instruct vs Qwen3 VL 8B Thinking: API price and specs
Qwen: Qwen3 VL 8B Thinking is 22% cheaper than Qwen 2.5 72B Instruct on input price.
Qwen 2.5 72B Instruct is 81% cheaper than Qwen: Qwen3 VL 8B Thinking on output price.
Qwen 2.5 72B Instruct is 48% cheaper than Qwen: Qwen3 VL 8B Thinking on a 1M-in / 200k-out workload.
Qwen: Qwen3 VL 8B Thinking has a 100% larger context than Qwen 2.5 72B Instruct.
| Field | Qwen 2.5 72B Instruct | Qwen: Qwen3 VL 8B Thinking |
|---|---|---|
| Lowest input, per 1M | $0.23 +27.8% | $0.18 |
| Median input, per 1M | $0.36 +100% | $0.18 |
| Lowest output, per 1M | $0.40 | $2.09 +422.5% |
| Median output, per 1M | $0.40 | $2.09 +422.5% |
| Lowest unit price | — | — |
| Cost of 1M in + 200k out | $0.31 | $0.60 +92.9% |
| Context | 131K +50% | 262K |
| Max output | 16K +50% | 33K |
| Providers | 11 | 2 +81.8% |
| Regions | US | — |
| Accepts | text | image, text |
| Returns | text | text |
| Type | Text models | Text models |
| Free option | no | no |
| Cheapest provider | requesty | deepinfra |
Your own workload: how many tokens the model gets (input) and gives back (output) - for example, in a month. The "Cost" row of the table above is recalculated for these numbers, for each model.
Computed from current offers each time the page opens; the best value of a row is bold, the percent is how far behind the best.
Providers selling both
Frequently asked questions
Which is cheaper, Qwen 2.5 72B Instruct or Qwen: Qwen3 VL 8B Thinking?
Qwen: Qwen3 VL 8B Thinking is 22% cheaper than Qwen 2.5 72B Instruct on input price. Qwen 2.5 72B Instruct is 81% cheaper than Qwen: Qwen3 VL 8B Thinking on output price. Qwen 2.5 72B Instruct is 48% cheaper than Qwen: Qwen3 VL 8B Thinking on a 1M-in / 200k-out workload.
Which has the larger context, Qwen 2.5 72B Instruct or Qwen: Qwen3 VL 8B Thinking?
Qwen 2.5 72B Instruct: 131K tokens; Qwen: Qwen3 VL 8B Thinking: 262K tokens.
Where can I use both Qwen 2.5 72B Instruct and Qwen: Qwen3 VL 8B Thinking?
deepinfra, kilo