About
Phi-4-mini-instruct is a lightweight open model built upon synthetic data and filtered publicly available websites - with a focus on high-quality, reasoning dense data. The model belongs to the Phi-4...
Related Models
Pricing
Input
$0.08
per 1M tokens
Output
$0.35
per 1M tokens
Blended
$0.15
per 1M tokens
Cheaper than 77% of models. Median price is $0.54/1M tokens.
Cost Calculator
Tokens per day1M
100K100M
Daily
$0.15
Monthly
$4.42
vs. Similar Models
Command-R+ (Apr '24)Q:0.0
$6.00+3968%
Olmo 3 7B InstructQ:-0.2
$0.13-15%
Llama 3.2 Instruct 11B (Vision)Q:+0.3
$0.24+66%
Jamba 1.5 MiniQ:-0.3
$0.25+69%
Performance
46
tokens/sec
Faster than 13% of models
0.89
seconds
Faster than 63% of models
0.89
seconds
Faster than 74% of models
Market Median
94 tok/s
51% slower
Median TTFT
1.10s
20% faster
Throughput/Dollar
311
tok/s per $/1M
Speed Comparison
Anthropic: Claude Opus 4.6
46 tok/s+0%
MiMo-V2.5-Pro (Non-reasoning)
46 tok/s+0%
Llama 3 Instruct 70B
47 tok/s+2%
Context Window
131K
tokens
Larger than 27% of models
Max Output
128K
tokens
98% of context
Benchmarks
MMLU-Pro
46.5%
GPQA Diamond
33.1%
HLE
4.2%
LiveCodeBench
12.6%
SciCode
10.8%
TerminalBench Hard
0.0%
MATH-500
69.6%
AIME
3.0%
AIME 2025
6.7%
IFBench
21.1%
Long Context Recall
13.7%
Tau2
8.2%
Market AverageTop Score