Fetching from the wire…
01
02
03
04
05
06
07
08
Models2026-08-06 · source-backed
128K context, 2.6B parameters, posting 51.87 AIME25, 59.17 IFBench, 56.88 BFCLv4 and 62.85 Claw-Eval, beating Gemma-4-E4B (8B) across all four and trailing Qwen3.5-9B narrowly. 220 tok/s on an M5 Max CPU, ~30 tok/s on phones, day-one support in llama.cpp, MLX, vLLM, SGLang and ONNX. Local tool-calling agents just became genuinely viable on consumer hardware, which changes the calculus for anything privacy-sensitive.
Each link below shares sources, entities, or timing with this story.
Gemma built by Google / Shared entities / Earlier coverage
Linked by a graph relationship (Gemma built by Google); both cover Gemma, Local, MLX, SGLang; earlier Gemma coverage from 2026-05-06.
Gemma built by Google / Shared entities / Same source domain / Earlier coverage
Linked by a graph relationship (Gemma built by Google); both cover E4B, Gemma, Qwen3; reported by the same outlet (huggingface.co).
Gemma built by Google / Shared entities / Shared topic / Earlier coverage
Linked by a graph relationship (Gemma built by Google); both cover E4B, Gemma; overlapping topics (beat, gemma).
Ollama supports Gemma / Shared entity: Qwen3 / Same source domain / Earlier coverage / Tension
Linked by a graph relationship (Ollama supports Gemma); both cover Qwen3; reported by the same outlet (huggingface.co).
Ollama supports Gemma / Shared entities / Earlier coverage
Linked by a graph relationship (Ollama supports Gemma); both cover Gemma, MLX; earlier Gemma coverage from 2026-07-12.
Ollama supports Gemma / Shared entity: Qwen3 / Shared topic / Earlier coverage
Linked by a graph relationship (Ollama supports Gemma); both cover Qwen3; overlapping topics (agent, became).
transformers supports ONNX / Shared entities / Earlier coverage
Linked by a graph relationship (transformers supports ONNX); both cover Gemma, MLX, SGLang; earlier Gemma coverage from 2026-06-04.
Ollama supports Gemma / Shared entity: Qwen3 / Earlier coverage / Tension
Linked by a graph relationship (Ollama supports Gemma); both cover Qwen3; earlier Qwen3 coverage from 2026-03-22.