AI Rate Card

Qwen3 235B-A22B Thinking 2507 vs Devstral Small 2505: API price and specs

Devstral Small 2505 is 55% cheaper than Qwen3 235B-A22B Thinking 2507 on input price.

Devstral Small 2505 is 80% cheaper than Qwen3 235B-A22B Thinking 2507 on output price.

Devstral Small 2505 is 68% cheaper than Qwen3 235B-A22B Thinking 2507 on a 1M-in / 200k-out workload.

Qwen3 235B-A22B Thinking 2507 has a 104.8% larger context than Devstral Small 2505.

FieldQwen3 235B-A22B Thinking 2507Devstral Small 2505
Lowest input, per 1M$0.11 +120%$0.05
Median input, per 1M$0.26 +325%$0.06
Lowest output, per 1M$0.60 +400%$0.12
Median output, per 1M$2.30 +945.5%$0.22
Lowest unit price——
Cost of 1M in + 200k out$0.23 +210.8%$0.074
Context262K128K +51.2%
Max output262K4K +98.4%
Providers133 +76.9%
RegionsUS—
Acceptstexttext
Returnstexttext
TypeText modelsText models
Free optionnono
Cheapest providerio-netio-net

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

deepinfra, io-net, trustedrouter

Frequently asked questions

Which is cheaper, Qwen3 235B-A22B Thinking 2507 or Devstral Small 2505?

Devstral Small 2505 is 55% cheaper than Qwen3 235B-A22B Thinking 2507 on input price. Devstral Small 2505 is 80% cheaper than Qwen3 235B-A22B Thinking 2507 on output price. Devstral Small 2505 is 68% cheaper than Qwen3 235B-A22B Thinking 2507 on a 1M-in / 200k-out workload.

Which has the larger context, Qwen3 235B-A22B Thinking 2507 or Devstral Small 2505?

Qwen3 235B-A22B Thinking 2507: 262K tokens; Devstral Small 2505: 128K tokens.

Where can I use both Qwen3 235B-A22B Thinking 2507 and Devstral Small 2505?

deepinfra, io-net, trustedrouter