Phi 4 pricing and routing profile
Phi 4 from Microsoft lists at $0.07 per million input tokens and $0.14 per million output tokens, with a 16,384-token context window. Prices from OpenRouter's catalogue on 2026-09-28.
Structured output
Cost per 1,000 requests
| Request shape | List cost |
|---|---|
| Classification or routing (600 tokens in, 10 out) | $0.0434 |
| Chat reply (2,000 in, 400 out) | $0.196 |
| Retrieval-augmented answer (8,000 in, 500 out) | $0.63 |
Before prompt caching. Try your own shape in the cost calculator.
Can Gimbix route to it first?
Not as a first-answer model through OpenRouter: no host returns token log probabilities for Phi 4. It can still be the model Gimbix falls back to when a cheaper model is unsure.
Who serves Phi 4
| Host | Input / output per 1M | Log probabilities | Uptime, last 30 min |
|---|---|---|---|
| DeepInfra | $0.07 / $0.14 | no | 100.0% |
A host that lists log probabilities may still omit them; Gimbix checks each one before relying on it.
Cheaper models that can answer first
Models priced below Phi 4 that expose a confidence signal. Whether they answer your traffic as well is what a Gimbix pilot measures.
| Model | Input / output per 1M | Cheaper on a chat reply |
|---|---|---|
| Llama 3.1 8B Instruct | $0.05 / $0.08 | 33% |
| Qwen3.7 Flash | $0.03 / $0.13 | 43% |
| Granite 4.0 Micro | $0.017 / $0.112 | 60% |
| gpt-oss-20b | $0.018 / $0.09 | 63% |
| Ling 3.0 Flash | $0.021 / $0.063 | 66% |
| Ling 3.0 Flash VL | $0.021 / $0.0616 | 66% |
Compare Phi 4
- Phi 4 vs Claude Opus 5.5
- Phi 4 vs Claude Sonnet 5
- Phi 4 vs Claude Fable 5.1
- Phi 4 vs Claude Opus 5
- Phi 4 vs GPT-6 Sol
- Phi 4 vs GPT-6 Astra
- Phi 4 vs GPT-5.6 Sol
- Phi 4 vs GPT-5.6 Terra
- Phi 4 vs GPT-5.5
- Phi 4 vs GPT-5.4
- Phi 4 vs Gemini 3.1 Pro Preview
- Phi 4 vs Gemini 3.8 Flash
- Phi 4 vs Grok 4.7
- Phi 4 vs Qwen3.8 Max (0902)
- Phi 4 vs Kimi K3
- Phi 4 vs GLM 5.3
- Phi 4 vs DeepSeek V4 Pro 0423
- Phi 4 vs Mistral Large 3 2512
- Phi 4 vs Nova Premier 1.0
- Phi 4 vs Command A
- Phi 4 vs GPT-6 Luna
- Phi 4 vs GPT-5.6 Luna
- Phi 4 vs GPT-5.4 Mini
- Phi 4 vs GPT-5.4 Nano
- Phi 4 vs GPT-4.1 Mini
- Phi 4 vs gpt-oss-120b
- Phi 4 vs gpt-oss-20b
- Phi 4 vs Claude Haiku 4.5
- Phi 4 vs Gemini 3.5 Flash Lite
- Phi 4 vs Gemma 4 31B
- Phi 4 vs DeepSeek V4.1 Flash
- Phi 4 vs DeepSeek V4 Flash 0731
- Phi 4 vs GLM 5.3 Flash
- Phi 4 vs Hy4 preview
- Phi 4 vs MiMo-V2.6-Flash
- Phi 4 vs Nemotron 3.5 Lightning
- Phi 4 vs Nemotron 3 Ultra
- Phi 4 vs Qwen3.8 Flash
- Phi 4 vs Qwen3.7 Flash
- Phi 4 vs Qwen3.8 27B
- Phi 4 vs MiniMax M3
- Phi 4 vs Kimi K2.6
- Phi 4 vs Mistral Small 4
- Phi 4 vs Ministral 3 8B 2512
- Phi 4 vs Llama 4 Maverick
- Phi 4 vs Llama 3.3 70B Instruct
- Phi 4 vs Nova Lite 1.0
- Phi 4 vs Nova Micro 1.0
- Phi 4 vs Command R7B (12-2024)
Questions
How much does Phi 4 cost?
Phi 4 lists at $0.07 per million input tokens and $0.14 per million output tokens on OpenRouter as of 2026-09-28. A typical chat reply of 2,000 input and 400 output tokens costs about $0.196 per 1,000 requests.
What is the context window of Phi 4?
16,384 tokens.
Can a cheaper model replace Phi 4?
Sometimes. Gimbix proves it on your own traffic before routing: the cheaper model must match Phi 4 on one half of your requests and make at most three answers worse on the other.