Back

Microsoft

Phi-4 Multimodal Instruct

Open weightsHugging Face
Intelligence
5.8
Rank 647 of 744 measured models
Output speed
16.6 t/s
Rank 399 of 403 measured models
Price/1M
$0.14
Rank 125 of 659 measured models
Context window
128K
Rank 773 of 1199 measured models

Phi-4 Multimodal Instruct scores 5.8 on the Artificial Analysis Intelligence Index, ranking 647th of 744 measured models. The closest models on this index are Gemini 1.0 Ultra (5.8), Mistral Large (Feb '24) (5.8), and NVIDIA Nemotron Nano 12B v2 VL (5.8). At $0.14 per million tokens (3:1 input/output blended price), it is cheaper than Mistral Large (Feb '24) ($6) and NVIDIA Nemotron Nano 12B v2 VL ($0.30). With an output speed of 16.6 t/s, it is slower than NVIDIA Nemotron Nano 12B v2 VL (206 t/s).

Artificial Analysis Indices
Intelligence5.8
Rank 647 of 744 measured models
Standard Benchmarks
MMLU Pro48.5%
Rank 303 of 349 measured models
GPQA Diamond31.5%
Rank 572 of 619 measured models
HLE5.0%
Rank 478 of 664 measured models
LiveCodeBench13.1%
Rank 299 of 343 measured models
MATH-50069.3%
Rank 146 of 202 measured models
AIME 20249.3%
Rank 136 of 194 measured models
Models at a similar level
The closest models on the Intelligence Index, one configuration per family.

Preferences and analytics

Rybbit: visits, performance, clicks, copies, forms and session replay (masked inputs), with your permission. Refusing does not limit the site.

Privacy

Allow navigation analytics?

With your permission, Rybbit measures visits, performance and interactions (clicks, copies, forms), including session replay, to improve BenchSift. Input values are masked in replays. Refusing does not limit the site.

Your choice is stored for 180 days. Change it at any time using “Privacy preferences” in the footer. Withdrawal reloads the page.

No choice stored: Rybbit is disabled.

Privacy