Skip to content
fedi.software

phi-3-medium-4k-instruct in fedi-index

The score in every slice, what it is made of, value, local run and history.

fedi-index · General

D 1,149 ±3

Percentile 18 · #239 · provisional

People prefer / Solves tasks

1,146 / —

Human votes vs benchmarks, Elo-eq

Value

-199

$0.2975 per 1M tokens, median of 1 providers

On your hardware

Closed weights: cloud only.

Score decomposition

D General 1,149 [1,146–1,152] Percentile 18 · #239 · components: 6 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena text/overall 1,138 ±2.61 1,138 ±3 26.5% 2026-10-02 —
LMArena text/hard_prompts 1,127 ±4.11 1,148 ±4 25.7% 2026-10-02 —
LMArena text/instruction_following 1,114 ±3.74 1,142 ±4 12.9% 2026-10-02 —
LMArena text/expert 1,108 ±9.08 1,156 ±7 11.0% 2026-10-02 —
LMArena text/multi_turn 1,088 ±5.58 1,116 ±5 6.1% 2026-10-02 —
LMArena text/longer_query 1,120 ±6.60 1,139 ±6 5.8% 2026-10-02 —

Weight is the share of the component in the score (the rest is the prior that pulls models with little data towards the population median). Methodology

D Coding 1,147 [1,139–1,155] Percentile 16 · #235 · components: 1 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena text/coding 1,130 ±5.19 1,141 ±5 89.8% 2026-10-02 —

Weight is the share of the component in the score (the rest is the prior that pulls models with little data towards the population median). Methodology

D Math 1,179 [1,170–1,188] Percentile 21 · #201 · components: 1 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena text/math 1,173 ±5.38 1,176 ±5 87.7% 2026-10-02 —

Weight is the share of the component in the score (the rest is the prior that pulls models with little data towards the population median). Methodology

D Writing 1,137 [1,129–1,145] Percentile 18 · #221 · components: 2 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena text/creative_writing 1,107 ±5.49 1,128 ±6 62.7% 2026-10-02 —
LMArena text/longer_query 1,120 ±6.60 1,139 ±6 30.6% 2026-10-02 —

Weight is the share of the component in the score (the rest is the prior that pulls models with little data towards the population median). Methodology

D Language: 繁體中文 1,146 [1,136–1,156] Percentile 11 · #199 · components: 1 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena text/chinese 1,108 ±6.55 1,142 ±5 93.5% 2026-10-02 —

Weight is the share of the component in the score (the rest is the prior that pulls models with little data towards the population median). Methodology

D Language: Deutsch 1,149 [1,132–1,166] Percentile 10 · #129 · components: 1 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena text/german 1,101 ±11 1,144 ±9 92.0% 2026-10-02 —

Weight is the share of the component in the score (the rest is the prior that pulls models with little data towards the population median). Methodology

D Language: English 1,148 [1,141–1,154] Percentile 20 · #219 · components: 1 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena text/english 1,169 ±3.36 1,144 ±4 93.8% 2026-10-02 —

Weight is the share of the component in the score (the rest is the prior that pulls models with little data towards the population median). Methodology

D Language: Español 1,126 [1,100–1,151] Percentile 4 · #143 · components: 1 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena text/spanish 1,095 ±16 1,116 ±15 89.1% 2026-10-02 —

Weight is the share of the component in the score (the rest is the prior that pulls models with little data towards the population median). Methodology

D Language: Français 1,130 [1,102–1,158] Percentile 4 · #129 · components: 1 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena text/french 1,109 ±18 1,120 ±16 88.2% 2026-10-02 —

Weight is the share of the component in the score (the rest is the prior that pulls models with little data towards the population median). Methodology

D Language: 日本語 1,163 [1,146–1,180] Percentile 8 · #118 · components: 1 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena text/japanese 1,042 ±12 1,159 ±9 92.1% 2026-10-02 —

Weight is the share of the component in the score (the rest is the prior that pulls models with little data towards the population median). Methodology

D Language: 한국어 1,105 [1,087–1,123] Percentile 3 · #132 · components: 1 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena text/korean 954 ±13 1,096 ±10 91.8% 2026-10-02 —

Weight is the share of the component in the score (the rest is the prior that pulls models with little data towards the population median). Methodology

D Language: Русский 1,179 [1,170–1,189] Percentile 12 · #206 · components: 1 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena text/russian 1,145 ±5.83 1,178 ±5 93.4% 2026-10-02 —

Weight is the share of the component in the score (the rest is the prior that pulls models with little data towards the population median). Methodology

History

fedi-index by month
The same in a table
Model 2026-10
phi-3-medium-4k-instruct 1,149 D

Badge

Show the model's fedi-index on your site or in a README. The badge updates itself; it links to this page.

fedi-index: phi-3-medium-4k-instruct

Data CC BY 4.0: keep the link to fedi.software.