model family

Phi 4

Published by microsoft. 12 members have been measured on real machines. Every number below was signed by the machine that produced it — none of it is from a model card.

what it takes, what it returns
text → text

Declared by the network’s catalog, not measured — a signature says what this family would be asked. Whether it answered is the signed capability flags on each member.

Builds this network knows of

Declared, not measured — this is what the catalog says exists and what it would cost a machine to hold. Nothing here is signed, and none of it claims the build works.

Phi 4 BASE
hf.co/bartowski/phi-4-GGUF:IQ2_XSIQ2_XSollama4.2 GiB on diskquantized by bartowski
hf.co/microsoft/phi-4-gguf:Q3_K_SQ3_K_Sollama6.1 GiB on diskpublished by microsoft
hf.co/bartowski/phi-4-GGUF:IQ4_XSIQ4_XSollama7.4 GiB on diskquantized by bartowski
hf.co/bartowski/phi-4-GGUF:Q4_K_MQ4_K_Mollama8.4 GiB on diskquantized by bartowski
hf.co/microsoft/phi-4-gguf:Q5_1Q5_1ollama10.3 GiB on diskpublished by microsoft
hf.co/microsoft/phi-4-gguf:BF16BF16ollama27.3 GiB on diskpublished by microsoft
Phi 4 MINI
hf.co/bartowski/microsoft_Phi-4-mini-instruct-GGUF:IQ2_MIQ2_Mollama1.4 GiB on diskquantized by bartowski
hf.co/bartowski/microsoft_Phi-4-mini-instruct-GGUF:Q3_K_SQ3_K_Sollama1.8 GiB on diskquantized by bartowski
hf.co/MaziyarPanahi/Phi-4-mini-instruct-GGUF:Q3_K_LQ3_K_Lollama2.1 GiB on diskderived by MaziyarPanahi
hf.co/MaziyarPanahi/Phi-4-mini-instruct-GGUF:Q4_K_MQ4_K_Mollama2.3 GiB on diskderived by MaziyarPanahi
hf.co/bartowski/microsoft_Phi-4-mini-instruct-GGUF:Q5_K_LQ5_K_Lollama2.8 GiB on diskquantized by bartowski
hf.co/unsloth/Phi-4-mini-instruct-GGUF:BF16BF16ollama7.2 GiB on diskquantized by unsloth
Phi 4 MINI-REASONING
hf.co/unsloth/Phi-4-mini-reasoning-GGUF:IQ1_SIQ1_Sollama1.1 GiB on diskquantized by unsloth
hf.co/unsloth/Phi-4-mini-reasoning-GGUF:Q2_K_LQ2_K_Lollama1.6 GiB on diskquantized by unsloth
hf.co/unsloth/Phi-4-mini-reasoning-GGUF:Q3_K_MQ3_K_Mollama2.0 GiB on diskquantized by unsloth
hf.co/lmstudio-community/Phi-4-mini-reasoning-GGUF:Q4_K_MQ4_K_Mollama2.3 GiB on diskquantized by lmstudio-community
hf.co/unsloth/Phi-4-mini-reasoning-GGUF:Q6_KQ6_Kollama2.9 GiB on diskquantized by unsloth
hf.co/unsloth/Phi-4-mini-reasoning-GGUF:BF16BF16ollama7.2 GiB on diskquantized by unsloth
Phi 4 REASONING-PLUS
hf.co/unsloth/Phi-4-reasoning-plus-GGUF:IQ1_SIQ1_Sollama3.3 GiB on diskquantized by unsloth
hf.co/unsloth/Phi-4-reasoning-plus-GGUF:Q2_K_LQ2_K_Lollama5.3 GiB on diskquantized by unsloth
hf.co/unsloth/Phi-4-reasoning-plus-GGUF:Q3_K_XLQ3_K_XLollama6.9 GiB on diskquantized by unsloth
hf.co/lmstudio-community/Phi-4-reasoning-plus-GGUF:Q4_K_MQ4_K_Mollama8.4 GiB on diskquantized by lmstudio-community
hf.co/unsloth/Phi-4-reasoning-plus-GGUF:Q6_KQ6_Kollama11.2 GiB on diskquantized by unsloth
hf.co/unsloth/Phi-4-reasoning-plus-GGUF:BF16BF16ollama27.3 GiB on diskquantized by unsloth
ask your own gram

Your concierge can answer this from the network rather than from this page. The ask carries the operator id, so it resolves to a capability and an engine — not to a name somebody has to recognise.

What has this network measured about the Phi 4 family (epn.family.phi-4), and can my gram run any of it? See /inferences/phi-4

open a gramx with this ask →epn.family.phi-4