This is could be bad news for OpenAI and Anthropic, given that most anyone with decent AI GPUs can now become their direct competitor, and the fact that Kimi K3 is open-weight also gives off the impression that "free" software is nearly as good, and far cheaper to run, than its proprietary competitors. It's worth noting that open-weight does not mean open-source; the training process and dataset are still Moonshot's special secret sauce.
Follow Tom's Hardware on Google News , or add us as a preferred source , to get our latest news, analysis, & reviews in your feeds.
Bruno Ferreira is a contributing writer for Tom's Hardware. He has decades of experience with PC hardware and assorted sundries, alongside a career as a developer. He's obsessed with detail and has a tendency to ramble on the topics he loves. When not doing that, he's usually playing games, or at live music shows and festivals. ","collapsible":{"enabled":true,"maxHeight":250,"readMoreText":"Read more","readLessText":"Read less"}}), "https://slice.vanilla.futurecdn.net/13-4-25/js/authorBio.js"); } else { console.error('%c FTE ','background: #9306F9; color: #ffffff','no lazy slice hydration function available'); } Bruno Ferreira Social Links Navigation Contributor Bruno Ferreira is a contributing writer for Tom's Hardware. He has decades of experience with PC hardware and assorted sundries, alongside a career as a developer. He's obsessed with detail and has a tendency to ramble on the topics he loves. When not doing that, he's usually playing games, or at live music shows and festivals.
thisisaname The quuestion many will ask is not cost or how easy it is to run (which is sort of cost again) but can it be trusted? Reply
alrighty_then They distilled American tech (Claude/Fable) then threw more parameters at it on American (Nvidia) hardware which was banned in China last I checked… This is the modus operandi, to duplicate and attempt to improve upon existing products, cutting corners as needed. I don't trust any claims until proven by many independent sources and of course they cannot be trusted with any real/sensitive data. Reply
bimo_tok thisisaname said: The quuestion many will ask is not cost or how easy it is to run (which is sort of cost again) but can it be trusted? It is very simple to test it if it is accurate or not. Please use your brain. Reply
bimo_tok alrighty_then said: They distilled American tech (Claude/Fable) then threw more parameters at it on American (Nvidia) hardware which was banned in China last I checked… This is the modus operandi, to duplicate and attempt to improve upon existing products, cutting corners as needed. I don't trust any claims until proven by many independent sources and of course they cannot be trusted with any real/sensitive data. "Microsoft tests Kimi K3 for Copilot in bid to cut AI costs by $600 million" Anti China propaganda detected! Reply
zsydeepsky thisisaname said: The quuestion many will ask is not cost or how easy it is to run (which is sort of cost again) but can it be trusted? I would argue: an open-source model with comprehensive tech reports & 3rd party use tests, seems infinitely more trustworthy than private models. Reply
Bizt alrighty_then said: They distilled American tech (Claude/Fable) then threw more parameters at it on American (Nvidia) hardware which was banned in China last I checked… This is the modus operandi, to duplicate and attempt to improve upon existing products, cutting corners as needed. I don't trust any claims until proven by many independent sources and of course they cannot be trusted with any real/sensitive data. Moonshot made k3 available 2 weeks after fable 5 was released. If moonshot distilled fable to build k3, they would have to do it, prepare infrastructure and get everything ready to meet the time table. K3 is a 2.8T model. It is impossibility to get everything ready in 14 days. Anyone with a brain, and a little knowledge about training llm would know better. Reply
thisisaname bimo_tok said: It is very simple to test it if it is accurate or not. Please use your brain. No these "AI"'s are anything but simple andvery hard to test, just look at how that AI escaped it sandbox and hacked another company! Reply
thisisaname zsydeepsky said: I would argue: an open-source model with comprehensive tech reports & 3rd party use tests, seems infinitely more trustworthy than private models. Could be but open does not automatical means it is better and are they fully open to where the weights come from? Reply
zsydeepsky thisisaname said: Could be but open does not automatical means it is better and are they fully open to where the weights come from? you seem to have totally ignored the " tech report " part I mentioned: https://github.com/MoonshotAI/Kimi-K3/blob/main/k3_tech_report.pdf Meanwhile, Anthropic purchases rare books, scans them, then destroys them into ashes to secure their "advantages": https://futurism.com/future-society/anthropic-destroying-books Reply
GreekChorus thisisaname said: The quuestion many will ask is not cost or how easy it is to run (which is sort of cost again) but can it be trusted? That question needs to be asked of all AI/LLM models, no matter what their source. Reply
Key considerations
- Investor positioning can change fast
- Volatility remains possible near catalysts
- Macro rates and liquidity can dominate flows
Reference reading
- https://www.tomshardware.com/tech-industry/artificial-intelligence/SPONSORED_LINK_URL
- https://www.tomshardware.com/tech-industry/artificial-intelligence/moonshot-ai-releases-weights-for-kimi-k3-firing-a-shot-across-the-bow-of-openai-and-anthropic-open-weight-model-performs-almost-as-well-as-frontier-models-while-being-2-3x-easier-to-run#main
- https://www.tomshardware.com/subscription
- At AI Summit, South Korea Outlines Its AI Future With NVIDIA and Partners
- Creality's multicolor K2 3D printer hits an all-time low price of $499 — save $200 today, with the perfect printer for jumping into the hobby
- Intel unveiled its iconic Core 2 Duo family 20 years ago — legendary chip dethroned AMD Athlon, restoring the chipmaker’s performance lead
- Daring coder gets Doom running with regular expressions at 180 seconds per frame, like playing 'correspondence chess with a shotgun' — nearly 14 million substit
- How to look better on webcam — tips for the average person to look better and more professional
Informational only. No financial advice. Do your own research.