Unverified50% confidenceFactExact time

Candidate small models include SmolLM2 (135M and 360M), Gemma 3 270M, LFM2.5-350M, Qwen2.5-0.5B, and Qwen3-0.6B, all of which have ONNX browser builds.

1

Sources

50%

Confidence

Medium-term (~90 days)

Relevance

10/13/2026

First Seen

Valid until: 1/11/2027

Sources

Running RAG in the Browser: Field Tests and Reflections on Small Language Models

Related Entities

SmolLM2model Gemma 3 270Mmodel LFM2.5-350Mmodel Qwen2.5-0.5Bmodel Qwen3-0.6Bmodel ONNXconcept

Related Claims

UnverifiedModels in the 3B–4B range, such as Qwen2.5 3B or Llama 3.2 3B, require only 2 to 2.5GB of VRAM under Q4 quantization.68% similar UnverifiedSome ONNX exports labeled q4 are larger than the q4f16 versions because the embedding table retains 32-bit precision; for example, SmolLM2-360M's q4 is 390MB and q4f16 is 275MB, and the two produce identical outputs.68% similar UnverifiedThe Qwen3.8-Flash-Next model supports an ultra-long context window of 262K tokens.66% similar

Cite This Claim

Stable URI

https://kongchang.com/claim/1013585

API

curl https://kongchang.com/api/v1/knowledge/claims/1013585

MCP

get_claim(id=1013585)