Probably the best model report this year: multiple infrastructure pieces cohering to fix computation and network to hold 3T parameters.
mr. TIM@timkellogg.me · last wk.You’ve waited long enough, the Kimi K3 open weights & full tech report are here! github.com/MoonshotAI/K...
![K KIMI K3: OPEN FRONTIER INTELLIGENCE
TECHNICAL REPORT OF KIMI K3
Kimi Team
ABSTRACT
parameters, native vision capabilities, and a 1-million-token context window. Kimi K3 is built on Kimi Delta Attention [63] and Attention Residuals [57], which improve information flow across sequence length and model depth. Together with Stable LatentMoE, which effectively activates 16 of 896 routed experts per token, and refined training and data recipes, these advances yield an proximately 2.5x improvement in overall scaling efficiency over Kimi K2 [581. Post-trainin highlights reinforcement learning across general, agentic, and coding domains and multiple reasoning-effort levels, enabling compositional generalization and robust long-horizon execution. At 2.8T scale, Kimi K3 is supported by infrastructure advances in multiple areas: algorithm-system co-design for KDA, perfectly balanced expert-parallel training with efficient memory management, million-token agentic RL with persistent rollout and sandbox states, and deployment innovations.
Extensive evaluations show that Kimi K3 achieves frontier-level performance across long-horizon coding, agentic, knowledge, reasoning, and vision tasks. While its overall performance still trails the most powerful proprietary models, namely Claude Fable 5 and GPT-5.6 Sol, Kimi K3 consistently outperforms other open and proprietary models evaluated in our suite. We release the full Kimi K3 model weights to facilitate future research and accelerate the broader deployment and adoption of frontier intelligence.'](https://cdn.bsky.app/img/feed_thumbnail/plain/did:plc:ckaz32jwl6t2cno6fmuw2nhn/bafkreicrbgvxize5b2uxilew4qvymuq5jvasfva4zp4i4bohjtkvxvvdv4)
In Nolan's The Odyssey, Sue establishes the laws governing food and hospitality.
Zum Lachen: Eine "Master Class" für MEP Assistenten. Aber auch seltsam, dass MEPs für der filzige European Internet Foundation weiterhin Räumlichkeiten im Parlament bieten und dabei mitmachen. www.internetforum.eu/events/event...
European Internet Forum - Cloud Technology
EIF’s mission is to help provide European political leadership for the development of European and multilateral public policies responsive to the political, economic and social challenges of the world...
internetforum.eu
2/ I'd be fine and would appreciate that "manufacturers" of goods that include works of "software" authors become subject to a CE certification process (permission to market): webcams, digital radios etc. Yet, the Cyber Resilience Act also covers software works as such.
My main concern of the Cyber Resilience Act is not the amateur clause for open source but that it also covers software as such. Software is protected as a work of literature under the Acquis Communautaire, thus the CRA introduces pre-censorship of works from software authors (permission to market).
The Saffron Industry in Iran Faces Challenges Due to the Climate Crisis www.onegreenplanet.org/environment/...
The Saffron Industry in Iran Faces Challenges Due to the Climate Crisis
The saffron industry is grappling with a disaster, as changing weather patterns and water shortages take a toll on the precious spice.
onegreenplanet.org
With the November GISTEMP data, the projected 2023 annual mean is certain to be new record, and clearly outside of the spread of the predictions I made at the beginning of the year - underlining the exceptional nature of how this year's anomalies unfolded.
