Unverified50% confidenceFactExact time
Gemini版本体系遵循双轴命名逻辑：主版本号代表架构或训练范式升级，规格后缀（Ultra、Pro、Flash、Nano）对应不同参数量级与部署场景
1
Sources
50%
Confidence
Long-term
Relevance
7/22/2026
First Seen
Sources
Related Claims
VerifiedThe Gemini series is divided into multiple tiers—Ultra, Pro, Flash, and Nano—corresponding to different compute requirements and deployment scenarios.82% similarUnverifiedGemini comes in three versions—Ultra, Pro, and Nano—targeting top-tier reasoning tasks, mainstream commercial scenarios, and on-device inference optimization respectively, replacing the previous Bard brand77% similarVerifiedGemini adopts a native multimodal architecture, jointly training the Transformer with mixed multimodal data from the pre-training stage, with tokens of different modalities processed together in the same sequence.75% similarUnverifiedSince its launch in late 2023, the Gemini model family has formed a three-tier product architecture: Ultra positioned as the highest-capability flagship, Pro as a general-purpose high-performance model, and Flash emphasizing ultra-fast inference and low-latency deployment.74% similarUnverifiedGemini 2.5 Flash adopts a Mixture of Experts (MoE) architecture, activating only a subset of parameters during inference.73% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/587654API
curl https://kongchang.com/api/v1/knowledge/claims/587654MCP
get_claim(id=587654)