Inkling Small is 2.1× cheaper per million tokens (blended). Qwen3.8 27B has the larger context window (1,000,000 vs 524,288 tokens, 1.9×).
Qwen3.8 27B vs Inkling Small
Live catalog fields. Best-in-row is highlighted.
Qwen3.8 27Bopen
Qwen · qwen__qwen3.8-27b
Inkling Smallopen
Thinkingmachines · thinkingmachines__inkling-small
Identity
| Field | qwen__qwen3.8-27b | thinkingmachines__inkling-small |
|---|---|---|
| Provider | Qwen | Thinkingmachines |
| Slug | qwen__qwen3.8-27b | thinkingmachines__inkling-small |
| Status | live | live |
| Open weights | Yes | Yes |
| License | — | — |
| HuggingFace | Qwen/Qwen3.8-27B | thinkingmachines/Inkling-Small |
Quality
| Field | qwen__qwen3.8-27b | thinkingmachines__inkling-small |
|---|---|---|
| Quality index (0–100) | 89.0 | — |
Cost
| Field | qwen__qwen3.8-27b | thinkingmachines__inkling-small |
|---|---|---|
| Input $ / 1M tokens | $0.42 | $0.45 |
| Cached input $ / 1M | $0.09 | $0.10 |
| Output $ / 1M tokens | $3.00 | $1.20 |
Context
| Field | qwen__qwen3.8-27b | thinkingmachines__inkling-small |
|---|---|---|
| Max context tokens | 1,000,000 | 524,288 |
| Max output tokens | 131,072 | 262,144 |
Modalities
| Field | qwen__qwen3.8-27b | thinkingmachines__inkling-small |
|---|---|---|
| Input | text, image, video | text, image, audio |
| Output | text | text |