Qwen3 235B A22B Instruct 2507 pricing
Qwen3 235B A22B Instruct 2507, Flagship-tier from Novita AI, costs $0.09 per 1M input tokens and $0.58 per 1M output tokens.
Input$0.09per 1M tokens
Output$0.58per 1M tokens
Context window128Ktokens
TierFlagshipNovita AI
Price your own workload against Qwen3 235B A22B Instruct 2507 in the calculator →
Specs and pricing
| Provider | Novita AI |
| Tier | Flagship |
| Context window | 128K tokens |
| Max output | — |
| Input price | $0.09 / 1M tokens |
| Output price | $0.58 / 1M tokens |
| Modalities | text |
Note. Serverless per-token pricing on Novita's model catalog; dedicated GPU deployments are billed per hour instead.
What it costs at realistic volumes
Assuming a 4:1 input:output token ratio, a common shape for chat and RAG-style workloads:
| Volume (4:1 in:out ratio) | Input cost | Output cost | Total / month |
|---|---|---|---|
| 1,000,000 in / 250,000 out | $0.09 | $0.14 | $0.23 |
| 10,000,000 in / 2,500,000 out | $0.90 | $1.45 | $2.35 |
| 100,000,000 in / 25,000,000 out | $9.00 | $14.50 | $23.50 |
Cheaper alternatives
Same tier (Flagship), a different provider, and a lower blended cost at the 4:1 ratio used above:
No same-tier model from another provider is cheaper than Qwen3 235B A22B Instruct 2507 at a 4:1 input:output ratio, as of 2026-08-30.