Phi 4
Microsoft Research developed Phi-4 as a 14-billion parameter model designed for complex reasoning tasks. It is built to operate efficiently in environments with limited memory while maintaining high performance for quick response requirements.
Phi 4 specs and pricing
- Provider: Microsoft
- Released: January 10, 2025
- Context window: 16K tokens
- Max output: 15K tokens
- Input: text
- Output: text
- Price: $0.07 input / $0.14 output per 1M tokens
What Phi 4 is good at
- Complex reasoning tasks
- Applications with limited memory
- Scenarios requiring quick response times
To-dos to try with Phi 4
- Analyze this logical puzzle and explain the step-by-step reasoning used to reach the solution.
- Summarize the core arguments in this technical report while maintaining a concise format.
- Draft a structured outline for a project plan based on these specific constraints.
FAQ
- Who developed Phi-4? Phi-4 was developed by Microsoft Research.
- What is the primary design focus of this model? The model is designed to excel at complex reasoning tasks while remaining efficient enough for use in memory-constrained environments.
- How much does Phi 4 cost? Phi 4 costs $0.07 per million input tokens and $0.14 per million output tokens through OpenRouter. In todo.is a typical agent step costs about 0.22 credits ($1 = 100 credits).
- What is the context window of Phi 4? Phi 4 can read up to 16K tokens at once and write up to 15K tokens in one reply.
- Can I use Phi 4 without coding? Yes. In todo.is you pick Phi 4 as your agent's model and write to-dos in plain words, on the web, the Android app, WhatsApp or Telegram. No API key needed: credits are included, or you can bring your own key.
JavaScript is required to use the todo.is app.