VQV Signal

ROBOTS & HARDWARE WATCH TECHNICAL

Strata 0.2.19 adds native Inkling 2 Mini support

According to a post on X by @coldniko, Strata 0.2.19 now runs Thinking Machines' Inkling 2 Mini natively, at 11,306 tokens/s prefill and 4,087 tokens/s output on an RTX 5090. The post also says Qwen3.8-Flash-Next gained a 40% speedup across supported hardware. The GitHub release confirms a new strata-inkling engine for the 26.4B-parameter looped model. The speed figures are the developer's own, and the post is described as mainly a video.

Source: Scobleizer News Pulse (X) · x.com Found via Robert Scoble’s Scobleizer News Pulse · a post on X, not independently verified Published 2026-11-22T14:45:00+00:00 Detected 2026-11-22T16:03:19+00:00
View original source

According to a post on X by @coldniko, Strata 0.2.19 now runs Thinking Machines' Inkling 2 Mini natively, at 11,306 tokens/s prefill and 4,087 tokens/s output on an RTX 5090. The post also says Qwen3.8-Flash-Next gained a 40% speedup across supported hardware. The GitHub release confirms a new strata-inkling engine for the 26.4B-parameter looped model. The speed figures are the developer's own, and the post is described as mainly a video.

AI-assisted summary based on the listed source.

@coldniko says Strata now runs Thinking Machines' Inkling 2 Mini at 11,306 tokens/s prefill and 4,087 tokens/s output on an RTX 5090, and that Qwen3.8-Flash-Next got a 40% speedup across supported hardware; 13 people in the lists shared it. The GitHub release confirms a new strata-inkling engine for the...

Developers who run open-source models on local hardware may find these claims useful if the speeds hold up in their own setups. The confirmed change is the new strata-inkling engine, so the speed numbers should be treated as the developer's claims until independent benchmarks appear. Watch for further release notes and third-party tests.

Hardware and robotics watchers may want to track whether this becomes a product, benchmark, or deployment signal.

Signal Strength 88% Technical label WATCH Public Interest 43 Category ROBOTS & HARDWARE Reader Depth TECHNICAL

Signal Strength reflects source quality, relevance, freshness and evidence. Public Interest helps organize discovery; it is not proof of truth.

Public Interest components
Recognizable Entity Score 0 Practical Impact Score 64 Novelty Interest Score 94 Consequence Score 18 Curiosity Score 32 Shareability Score 57

VQV surfaced this signal because it is recent, relevant to Open Source LLMs, connected to Scobleizer News Pulse (X).