Phi-4-multimodal
5.6B tri-modal model handling text, vision, and audio together.
About
Microsoft's Phi-4-multimodal is a 5.6B-parameter model that processes text, images, and audio in a single model, built on the Phi-4-mini backbone with a Mixture-of-LoRAs technique. It holds a 128K-token context, tops the Hugging Face OpenASR leaderboard at 6.14% WER, and is released under the MIT license.
Vision input is English-focused and audio is limited to 40-second clips (30 minutes for summarization), so it is not a general audio workhorse.
Specs
Context window
Pricing
Free quota
API access
Weights
Released
Benchmarks
6.14% WER on OpenASR leaderboard (#1, Microsoft-reported); beats Whisper V3 on speech recognition (provider claim).
Provider-reported unless a source says otherwise. Benchmarks are marketing-adjacent by nature — treat them as directional.
Use cases
Trust score — 8/10
Scored 8/10 (verified tier, metrics-v1). Evidence: MIT license, 128K context, #1 on OpenASR leaderboard at release; tri-modal single model is genuinely useful for edge devices; vision is English-focused; audio clip limits apply. License: MIT per Microsoft's model card and release coverage..
How the trust score is built
Full methodology →- Adoption · stars/downloads, log-scaled vs peers
- 8.0/10
- Maintenance · recency, releases, issue load
- 7.0/10
- Openness · license + docs
- 9.0/10
- Manual review · sanity read + red-flag check
- 8.0/10
- · MIT license, 128K context, #1 on OpenASR leaderboard at release
- · tri-modal single model is genuinely useful for edge devices
- · vision is English-focused; audio clip limits apply