Model
Model explorer

Phi-3.5-vision (4.2B)

OPEN
Microsoft · Phi-3.5 family · released Aug 20, 2024

Multimodal Phi with 4.2B params and 128K context; adds multi-frame image understanding over Phi-3-vision.

ReasoningCodingVisionFunction callingTool useAgentic
532.4
Elo · unrated
Parameters
4.2B
Active params
4.2B (dense)
Context
128K tokens
Architecture
Dense decoder-only Transformer + vision encoder/connector (Phi-3 Mini backbone)
License
MIT
Languages
API price (in/out)
$0.13 / $0.52
Modalities
text · vision
Benchmark results
Bar shows position within the tracked field; marker = field best
AI2DVision78.1%#37
best: Molmo 72B · 96.3%
ChartQAVision81.8%#27
best: MiniMax-VL-01 · 91.7%
MathVistaVision43.9%#58
best: Seed 2.1 Pro · 90.7%
MMBenchVision81.9%#1
best: this model · 81.9%
MMMUVision43.0%#103
best: Claude Fable 5 · 89.3%
TextVQAVision72.0%#21
best: Molmo 2 8B · 85.7%
Video-MMEVision50.8%#24
best: Seed 2.1 Pro · 89.2%
Run it locally
VRAM @ Q4
4 GB
VRAM @ FP16
9 GB
Fits on (Q4)
RTX 3060 12GBRTX 4070 Ti 16GBRTX 3090 24GBRTX 4090 24GBRTX 5090 32GBM4 Pro 48GBM3 Max 128GBM3 Ultra 512GBA100 80GBH100 80GBH200 141GBB200 192GB
Throughput data unavailable.
Quantizations
GGUF Q4 · MLX 4-bit
API price $0.13/$0.52 · each benchmark row carries its own source badge (see methodology)