Qwen 3 235 B A22 B Thinking 2507 pricing
Source: openrouter · updated 2026-10-27
Short answer: Qwen 3 235 B A22 B Thinking 2507 costs $0.450 per million input tokens and $3.50 per million output tokens. That is a 3:1 blended rate of $1.21 per million tokens, ranking it #358 of 704 tracked models by input price.
Price breakdown
| Metric | USD |
|---|---|
| Input, per 1M tokens | $0.450 |
| Output, per 1M tokens | $3.50 |
| Output รท input multiple | 7.78× |
| 3:1 blended, per 1M tokens | $1.21 |
| Context window | 128,000 tokens |
What a typical workload costs
| Monthly volume | Input cost | Output cost | Total | 1M in / 250K out | $0.450 | $0.875 | $1.32 |
|---|---|---|---|---|---|---|---|
| 10M in / 2.5M out | $4.50 | $8.75 | $13.25 | ||||
| 100M in / 25M out | $45.00 | $87.50 | $132 |
Other qwen models
| Model | Input | Output |
|---|---|---|
| Qwen 3.7 Flash | $0.030 | $0.130 |
| Qwen 3.5 Flash 02 23 | $0.065 | $0.260 |
| Qwen 3 Coder 30 B A3 B Instruct | $0.070 | $0.280 |
| Qwen 3 32 B | $0.080 | $0.280 |
| Qwen 3 Next 80 B A3 B Instruct | $0.090 | $1.10 |
Compare against
- Qwen 3 235 B A22 B Thinking 2507 vs Qwen 3 Coder 480 B A35 B Instruct Fp 8
- Qwen 3 235 B A22 B Thinking 2507 vs Mimo V2.6 Pro
Cheaper alternatives
Next cheaper option: Qwq Plus at $0.800 per 1M input tokens. Next step up: Grok Build 0.1 at $1.00 per 1M input tokens.