A smaller, instruction-tuned variant of Phi 4 designed to run efficiently on resource-constrained hardware while handling a 128K token context.
- Strengths
- Delivers practical instruction-following performance with a large context window and low resource overhead compared to larger models.
- Best for
- Applications requiring reasonable quality responses on edge devices, mobile, or systems with limited compute — a step down from the full Phi 4 in capability but significantly lighter in resource demands.
- Limitations
- Lacks the complex reasoning depth of the larger Phi 4, and is not suited for tasks requiring the parameter scale or mixture-of-experts approach of WizardLM-2 8x22B.
Input / 1M
$0.08
Output / 1M
$0.35
Cached input / 1M
$0.08
Context window
128K
Price history
Snapshots
| Effective | Input | Output | Cached in | Note | Source |
|---|---|---|---|---|---|
| 11 Jun 2026 | $0.08 | $0.35 | $0.08 | Imported from OpenRouter | openrouter.ai |