Side-by-side comparison

Pick up to 4 models. Differences are what matter for picking.

Models
Spec
Google
Gemini 2.5 Flash with Native Audio (Live API)
Verified 2026-05-08
DescriptionReal-time speech-in / speech-out conversational model powering Gemini Live. Native audio with affective dialog and tool use.
Released2025-06-17
Knowledge cutoff2025-01
Context window1.0M
Max output8K
Input $/1M
Output $/1M
Speed tierFast
Reasoning4/5
Coding3/5
Multilingual5/5
Modalities in
Text
Audio In
Vision
Video In
Modalities out
Text
Audio Out
Tool useYes
JSON modeYes
Open weightsNo
Best use cases
Voice
Chat
Agents
Best personas
Consumer
Developer Tools
Startup
Weak at
  • · batch / non-realtime workloads
  • · deepest reasoning
Compliance
HIPAA-eligibleSOC 2GDPRISO 27001FedRAMPAvailable via Vertex AI under Google Cloud compliance umbrella.
Data policyVertex AI / paid Gemini API: prompts not used to train models.
Sources
Tip: hop over to the Recommender to describe a problem and have AI suggest the best fit from this dataset.