Phi-4 Chat: pricing, context window and benchmarks
Phi-4 Chat is a multimodal model from Microsoft. It offers a 65,536-token context window, costs $3.66 per million input tokens and $14.24 per million output tokens, and averages 74.7 across the index's benchmark suite.
Pricing
Phi-4 Chat costs $3.66 per million input tokens and $14.24 per million output tokens. A blended 3:1 input-to-output mix works out to about $6.30 per million tokens. Cached input is listed at $0.53 per million tokens.
| Item | Price | Notes |
|---|---|---|
| Input | $3.66 | per 1M tokens |
| Output | $14.24 | per 1M tokens |
| Cached input | $0.53 | per 1M tokens |
| Blended 3:1 | $6.30 | estimate per 1M tokens |
| 1M in + 1M out | $17.90 | estimate |
Benchmarks
Scores below are modelled estimates on a 0–100 scale. Use them to compare relative position, not as official vendor results.
| Benchmark | Score | Rank | Percentile | Relative |
|---|---|---|---|---|
| MMLU-Pro | 78.4 | 230 | 62th | |
| GPQA Diamond | 70.1 | 258 | 57th | |
| SWE-bench Verified | 62.2 | 278 | 54th | |
| HumanEval | 80.7 | 217 | 64th | |
| MATH-500 | 74.9 | 330 | 45th | |
| AIME 2025 | 67.2 | 309 | 49th | |
| LiveCodeBench | 68.9 | 210 | 65th | |
| IFEval | 84.0 | 214 | 64th | |
| TruthfulQA | 79.1 | 212 | 65th | |
| HellaSwag | 86.4 | 200 | 67th | |
| MMMU | 70.2 | 294 | 51th | |
| Arena Elo | 1372.7 | 309 | 49th |
What Phi-4 Chat is good for
Phi-4 Chat is an open-weight model, so it can be self-hosted or run through a third-party inference provider. That matters for data residency, fine-tuning and long-run cost control. At $3.66 per million input tokens (hosted list price) and a 65,536-token context window, it is a practical option for teams that want more control than a closed API allows.
Strengths
- Input price of $3.66 per million tokens is at or below the index median.
- Open weights make self-hosting, fine-tuning and data-residency review possible.
- Accepts more than text, including image input where the provider supports it.
Trade-offs
- No standout weakness in the indexed specifications, but real-world behaviour still needs testing.
How it compares
| Model | Provider | Context | Input / 1M | Avg score |
|---|---|---|---|---|
| Llama 3 Modal | Modal | 32768 | $3.82 | 74.9 |
| Llama 3 Anyscale | Anyscale | 65536 | $3.94 | 74.7 |
| Llama 3 Lambda Chat | Lambda | 32768 | $3.73 | 74.7 |
| Llama 3 Replicate Chat | Replicate | 200000 | $3.74 | 74.7 |
| Phi-4 Chat | Microsoft | 65536 | $3.66 | 74.7 |
| Llama 3 Anyscale Chat | Anyscale | 65536 | $3.67 | 74.6 |
FAQ
How much does Phi-4 Chat cost?
Phi-4 Chat is listed at $3.66 per million input tokens and $14.24 per million output tokens. Cached input is $0.53 per million tokens. A blended 3:1 input-to-output mix is about $6.30 per million tokens.
What is the context window of Phi-4 Chat?
Phi-4 Chat has a context window of 65,536 tokens and can emit up to 8,192 output tokens per request in this index.
Is Phi-4 Chat open weights?
Yes. Phi-4 Chat is listed as an open-weight model under a open weights licence.
How does Phi-4 Chat score on benchmarks?
Phi-4 Chat averages 74.7 across the scored benchmarks in this index, ranking #262 of 600 models. Its strongest relative result is Arena Elo at 1372.7.
What is Phi-4 Chat best for?
Phi-4 Chat is an open-weight model, so it can be self-hosted or run through a third-party inference provider. That matters for data residency, fine-tuning and long-run cost control. At $3.66 per million input tokens (hosted list price) and a 65,536-token context window, it is a practical option for teams that want more control than a closed API allows.