← Leaderboard

Phi-3.5-mini-instruct 4bit MLX

Tested Rank #33 of 33 · 2/8 GAUNTLET progress
Generalist — 31.5 / 100 G Agentic — pending A Understanding — pending U Needle — pending N Thinking — pending T Live — pending L Engineering — pending E Throughput — 67.5 / 100 T

Hollow markers & dashed spokes: axis not yet scored

G
31.5
A
U
N
T
n/d
L
E
T
67.5

Specification

Parameters
3.8B
Architecture
phi3
Size on disk
2 GB
Quantization
4bit
Format
MLX
Reasoning (CoT)
No
Internal ID
M17
Mean speed
110.2 tok/s across suites
Stall census
0 stalls in 13 observed tests (0.0%)
Model card
huggingface.co

Suite results

Suite Score Avg / 20 Tests tok/s
General capability (13-task real-workload suite) 1a 81.9 / 260 6.3 13/13 110.2

Each test is LLM-judged 0–20 against a fixed rubric; suite max = tests × 20. Where fewer tests ran than the suite total, unrun tests count as zero toward the axis score. See methodology.