readable AI benchmarks (simplified)
Microsoft
Phi-4
Phi-4 has a Quality Score of 15.6, ranking 488th among 571 scored models. Per 1M tokens, pricing is $0.125 input and $0.500 output. It has a 16k token context window. Its Reliability Score is 22.1, ranking 450th among 573 scored models, below average (total average: 36.5).
Quality Score15.6coverage 100.0%: AA Intelligence and AA-Omniscience outcomes available
Value Scorenot available
Factual reliability22.1
Cache Discountnot available
Model specification
- Reasoning
- non-reasoning
- Input modalities
- none
- Output modalities
- none
- Context window
- 16k tokens
- Weights
- open weights
Token prices USD per 1M tokens
- Input
- $0.125
- Output
- $0.500
- Cached Input
- not available
- Cache write
- not available
- Cost per task
- not available
Capability
- Intelligence
- 5.9
- Coding
- not available
- Agentic
- not available
- Omniscience
- -55.7
- Correct
- 14.1%
- Blended price
- not available
Answer outcomes
Fully graded outcomes (Correct + Incorrect): 83.8%
Correct 14.1%Incorrect 69.8%Partial / not attempted 16.2%
Artificial Analysis benchmarks
- Humanity’s Last Exam
- 4%
- GPQA Diamond
- 57%
- CritPt
- 0%
- AA-Omniscience
- 22%
- AA-LCR
- 0%
Similar models
- Phi-4 Mini InstructMicrosoftQuality Score 14.6Value Score not available
- Phi-3 Mini Instruct 3.8BMicrosoftQuality Score not availableValue Score not available
- Phi-4 Multimodal InstructMicrosoftQuality Score not availableValue Score not available
- Llama 3 Instruct 70BMetaQuality Score 15.6Value Score not available
- Jamba 1.7 LargeAI21 LabsQuality Score 15.9Value Score not available
- LFM2 2.6BLiquid AIQuality Score 15.5Value Score not available