Phi 4 — Currently unavailable
Microsoft Research Phi-4 is designed to perform well in complex reasoning tasks and can operate efficiently in situations with limited memory or where quick responses are needed. At 14 billion...
Technical specifications
Independent benchmarks
Scores below come from Artificial Analysis, an independent third party. BazaarLink does not participate in the testing.
Intelligence 6
Coding
Math 18
MMLU 71
GPQA 57
This modelCategory leader
Overall intelligence
#543out of 638 models
Aggregated across academic benchmarks
Time to first token
#178out of 329 models
AA 全球 p50：2.54 秒
Intelligence
6
Mid-tierThis model6
Category leader53
Median12
Math
18
—This model18
Category leader99
Median53
MMLU Pro
71%
—This model71%
Category leader90%
Median75%
GPQA
57%
—This model57%
Category leader94%
Median69%
LiveCodeBench
23%
—This model23%
Category leader42%
Median42%
HLE
4%
—This model4%
Category leader53%
Median7%
SciCode
26%
—This model26%
Category leader60%
Median34%
IFBench
24%
—This model24%
Category leader83%
Median44%
τ²-Bench Telecom
0%
—This model0%
Category leader99%
Median46%
AA-LCR
0%
—This model0%
Category leader76%
Median40%
Terminal-Bench Hard
4%
—This model4%
Category leader66%
Median14%
Top 5 — CodingCoding Index leaderboard
2Claude Fable 5.1 (Adaptive Reasoning, Xhigh Effort, Default Fallback)
?/claude-fable-5-1-xhigh81$10 / $503Claude Fable 5.1 (Adaptive Reasoning, High Effort, Default Fallback)
?/claude-fable-5-1-high79$10 / $504GPT-5.6 Sol (xhigh)
openai/gpt-5-6-sol-xhigh78$4 / $20AASource: Artificial Analysis · Independent third-party evaluation; not influenced by BazaarLinkView full evaluation on AA →
Frequently Asked Questions
Can I still use Phi 4 through BazaarLink?
No. Phi 4 is currently unavailable through BazaarLink. See the active alternatives linked on this page.
Relay integrity checks
No integrity-check data is available for this model yet.
BazaarLink Probe verifies endpoints that claim to serve this model and flags behavior that differs from the reference baseline.
View the full integrity report →Related models and links
Related models and links