CTO and co-Founder @TheDataCity. Did data @OpenInnovates and @imactivate. I block anons for even the slightest rudeness.
Leeds, Yorkshire, UK
Joined March 2010
- Tweets238K
- Following1.9K
- Followers36.8K
- Likes11.3K
Tom Forth retweeted
I must say, it was difficult hearing the hurt in Gary Neville's voice at the end of that game, so I turned up the volume on the TV and could hear it much better.
Tom Forth retweeted
Town over the last 10 years.
A change for the better or are we thinking it should have been left as it was?
Manchester
📸 | @Renaker_
Tom Forth retweeted
1969: Beswick: Where City's stadium is now...Those long winding empty roads with so much wasteland. My gran had a pub round there mid 70's, vaguely remember! Different world, different Manchester.
Tom Forth retweeted
Higher score than Fable 5.1 and Astra 6, wild stuff.
Kimi K3.1 Pro launches at 4th on the Artificial Analysis Intelligence Index scoring 55, making large gains in agentic performance; it will be the leading open-weights model once its weights are published
@Kimi_Moonshot's new reasoning model gains 11 points over Kimi K3 (max, 44). It trails only Claude Opus 5.5 (max with fallback, 58), Claude Sonnet 5.5 (max with fallback, 56), and Muse Wonder Full (max, 56) on the Intelligence Index.
Once the weights for K3.1 Pro are released it will be the highest-scoring open weights model and the gap between the leading open weights and proprietary models on the Intelligence Index will narrow from 12 points to 3.
Key results:
➤ Large gains in agentic performance: Kimi K3.1 Pro scores 3rd on GDPval-AA with 1,846 Elo, up 312 Elo from Kimi K3, behind only Claude Opus 5.5 (1,866 Elo) and Muse Wonder Full (1,857 Elo). Kimi K3.1 Pro also achieves the highest score of any model on Terminal-Bench 4.0. K3.1 Pro scores 67%, up from 13% for Kimi K3 and ahead of Claude Sonnet 5.5 (64%). AutomationBench-AA, which tests agentic SaaS workflows, rises from 58% to 73%.
➤ Gains in other evals are more moderate: Humanity's Last Exam rises from 47% to 52%, CritPt from 23% to 25%, and SciCode from 59% to 61%. AA-Omniscience Index rises from 20 to 30, as the hallucination rate falls from 53% to 44% and accuracy rises from 48% to 51%.
➤ Cost per Task roughly equal to Kimi K3 at $1.98 per Task: K3.1 Pro uses 55K output tokens per Intelligence Index task, 14% more than Kimi K3 at 48K, while its first-party token prices are 33% lower.
Additional model details:
➤ Context window: 1M tokens.
➤ Size: 2.8T total, 104B active parameters
➤ Input modalities: Text and image
➤ Weights and license: Weights currently not published, Moonshot is expected to release the weights soon after launch
➤ Pricing: $2/$10/$0.20 per 1M input/output/cache hit tokens, 33% lower than Kimi K3 on all three rates.
Congratulations to the team at @Kimi_Moonshot on the release!
Tom Forth retweeted
Kimi K3.1 Pro launches at 4th on the Artificial Analysis Intelligence Index scoring 55, making large gains in agentic performance; it will be the leading open-weights model once its weights are published
@Kimi_Moonshot's new reasoning model gains 11 points over Kimi K3 (max, 44). It trails only Claude Opus 5.5 (max with fallback, 58), Claude Sonnet 5.5 (max with fallback, 56), and Muse Wonder Full (max, 56) on the Intelligence Index.
Once the weights for K3.1 Pro are released it will be the highest-scoring open weights model and the gap between the leading open weights and proprietary models on the Intelligence Index will narrow from 12 points to 3.
Key results:
➤ Large gains in agentic performance: Kimi K3.1 Pro scores 3rd on GDPval-AA with 1,846 Elo, up 312 Elo from Kimi K3, behind only Claude Opus 5.5 (1,866 Elo) and Muse Wonder Full (1,857 Elo). Kimi K3.1 Pro also achieves the highest score of any model on Terminal-Bench 4.0. K3.1 Pro scores 67%, up from 13% for Kimi K3 and ahead of Claude Sonnet 5.5 (64%). AutomationBench-AA, which tests agentic SaaS workflows, rises from 58% to 73%.
➤ Gains in other evals are more moderate: Humanity's Last Exam rises from 47% to 52%, CritPt from 23% to 25%, and SciCode from 59% to 61%. AA-Omniscience Index rises from 20 to 30, as the hallucination rate falls from 53% to 44% and accuracy rises from 48% to 51%.
➤ Cost per Task roughly equal to Kimi K3 at $1.98 per Task: K3.1 Pro uses 55K output tokens per Intelligence Index task, 14% more than Kimi K3 at 48K, while its first-party token prices are 33% lower.
Additional model details:
➤ Context window: 1M tokens.
➤ Size: 2.8T total, 104B active parameters
➤ Input modalities: Text and image
➤ Weights and license: Weights currently not published, Moonshot is expected to release the weights soon after launch
➤ Pricing: $2/$10/$0.20 per 1M input/output/cache hit tokens, 33% lower than Kimi K3 on all three rates.
Congratulations to the team at @Kimi_Moonshot on the release!
Tom Forth retweeted
The best one
Just Eat apologises for AI food mistakes after social media ridicule bbc.in/4jAIj9J
Tom Forth retweeted
Oh my goodness this has to be right.
Instead of generating a random 256-bit number they generated a random number between 0 and 256.
That’s how they got 4, 66, 155, 156, 230 and 237.
Tom Forth retweeted
The Data City Summit is back in Leeds on the of 12th November 🚀
Hear how people are using real-time data to find the frontier companies official statistics miss.
Talks, lunch and drinks are on us.
Register your ticket: hubs.ly/Q04B_b-c0
Tom Forth retweeted
London bus fares are massively subsidised by tube fares and the taxpayer. In some contexts, like buses that feed tubes or rail, this can be defended, but a large fraction of buses just duplicate rail, albeit at a cost for everyone else. One of the many drains on investment.
Tom Forth retweeted
One of the things I love about my role as a Minister in Number 10 North is that I get to work on progressing projects across the country to improve our economy.
A bonus is that sometimes the boss lets me hold meetings in Number 10 Downing Street!
So it was good to meet earlier with Peter Lowy, Director of Leeds United, to discuss how we can work with @TracyBrabin on the big regeneration planned at Elland Road.
Yes, we discussed the tram too.
Tom Forth retweeted
What effect has this weekend’s mob violence had on the polling for the 2027 🇫🇷 presidential election?
Answer: a big boost for Marine Le Pen
She is now polling at 34-38% in the first round
Mélenchon in strongest position to secure the other run-off place
source: @IfopOpinion
Tom Forth retweeted
Learning from this discourse that the whole USA is in a prisoner’s dilemma where they have 3% interchange fees that get passed through to the cost of goods so they can all make a hobby out of getting some of it back.
Tom Forth retweeted
It's convenient the dominant content style became fast-cut clip sequences just in time for ai that can only generate short clips. Almost enough to make you believe in the hyperobject reaching backwards through time to effect its own creation
I came to realize people liked the french #slopcore without having the lyrics 😅 sorry!
So here is a version with english subtitles!
Tom Forth retweeted
Encore plus héroïque avec le contexte : ce chef gaulois était en congé et s’est rendu à son magasin en s’inquiétant pour la sécurité de ses employés quand il a appris que la manif dégénérait
