Community MLX Challenge Pushes Gemma 4 26B to 151% Faster on Mac
TheMoonMidas · x · 2026-09-05
The mlx.fast community challenge for Gemma 4 26B A4B shows the model now runs 151.4% faster on Apple Silicon Macs than at launch, after 138 promoted submissions from 37 solvers.
- Official scoring compares baseline and candidate on the same Mac, combining gains as prefill^0.25 · decode^0.75, so decode speed carries three quarters of the score
- The current record was set by a Claude Opus-generated optimization, with top-5 entries from Claude, Gemini 3.8 Flash and GPT-5.6 Sol
- Gaps between adjacent ranks have narrowed to 0.1%, signaling diminishing low-hanging fruit
More from Infra
- Hermes adds a local backend with Unsloth UD-Q4 quants for DeepSeek-V4-Flash and Qwen models — maximelabonne · 2026-09-05
- AI Now on data center boom: community pushback and 'they won't build them where they live' — AINowInstitute · 2026-09-05
- Gemma 4 Runs 151.4% Faster on Mac via Community MLX Inference Optimization — gajesh · 2026-09-05
- SGLang's Breakable CUDA Graph speeds prefill graph building by 3.8–5.2x — ying11231 · 2026-09-05
- SemiAnalysis: OpenAI's ASIC program is leverage — Altman wins even if the chip loses — MarvinTBaumann · 2026-09-05
- 2027 will be peak year of AI compute constraint; relief arrives in 2028, analyst argues — BenBajarin · 2026-09-05