Qwen 3.6 27 B Fp 8 pricing
Source: litellm · updated 2026-10-28
Short answer: Qwen 3.6 27 B Fp 8 costs $0.320 per million input tokens and $3.20 per million output tokens. That is a 3:1 blended rate of $1.04 per million tokens, ranking it #338 of 713 tracked models by input price.
Price breakdown
| Metric | USD |
|---|---|
| Input, per 1M tokens | $0.320 |
| Output, per 1M tokens | $3.20 |
| Output รท input multiple | 10.00× |
| 3:1 blended, per 1M tokens | $1.04 |
| Context window | 262,144 tokens |
What a typical workload costs
| Monthly volume | Input cost | Output cost | Total | 1M in / 250K out | $0.320 | $0.800 | $1.12 |
|---|---|---|---|---|---|---|---|
| 10M in / 2.5M out | $3.20 | $8.00 | $11.20 | ||||
| 100M in / 25M out | $32.00 | $80.00 | $112 |
Other tensormesh models
| Model | Input | Output |
|---|---|---|
| Qwen 3 Coder 480 B A35 B Instruct Fp 8 | $0.450 | $1.80 |
| Qwen 3.5 397 B A17 B Fp 8 | $0.600 | $3.60 |
| GLM 5.1 Nvfp 4 Mtp | $1.40 | $4.40 |
Compare against
Cheaper alternatives
Next cheaper option: Mistral Large 4 at $0.680 per 1M input tokens. Next step up: GPT Audio Mini at $0.600 per 1M input tokens.