Lumina
@LuminaBench
• AI News and Benchmarks
• Our Discord -
• Tips/Contact - Luminainfo@proton.me
668 Following 14.1K Followers
When Fable 5.5 drops it might be over
OpenAI just can't catch a break
Anthropic are launching Claude Duet
This looks like their version of Grok bot / OpenAI's dots from what I can see
Anthropic are launching Claude Duet
This looks like their version of Grok bot / OpenAI's dots from what I can see
First impressions of Gemini 4 Carbon from arena testing:
It's extremely strong at 3D tasks / games, a massive improvement is that it's no longer lazy, it understands what you want takes its time and does it, instead of just trying to be quick like Argon
One negative is Frontend design / taste, it's still average and lacks what Anthropic models have
Show more
🚨 Kimi K3.1 is now getting prepped in Kimi Code
There is a model switching warning that now includes Kimi K3.1 if you change model mid session
No exact time of launch but it won't be far away now
Show more
Gemini 4 Carbon is incredible at making games
Checkpoint 1 is testing in arena currently and this is one of the cleanest outputs I have got from the Super Mario prompt
Gen time was quite high like 30 minutes for this but it's no longer lazy like Argon, it just gets what you want and does it perfectly first try
Show more
Gemini 4 Carbon's SVG Xbox Controller
This is with absolutely no tracing allowed or search allowed as I prompted it to not cheat
Gemini 4 Carbon (SVG Pelican on Bicycle)
Few things to point out, the UI isn't very good, and it's also incredibly slow
Otherwise it feels good an a vast array of tasks I have tried it on
Show more
Gemini 4 Carbon (SVG Pelican on Bicycle)
Few things to point out, the UI isn't very good, and it's also incredibly slow
Otherwise it feels good an a vast array of tasks I have tried it on
Show more
Gemini 4 Carbon Voxel Pagoda
Carbon has been in arena since yesterday, so I will show off some outputs I got over today
It's a noticeable improvement over Argon but is incredibly slow, like 2-4x slower depending on task
Show more
Gemini 4 Carbon (Voxel Pagoda) vs Frontier Models
These are some frontier models against it for reference, it has noticeably higher quality than Argon does, as i said it's a lot slower though
Show more
Gemini 4 Carbon Voxel Pagoda
Carbon has been in arena since yesterday, so I will show off some outputs I got over today
It's a noticeable improvement over Argon but is incredibly slow, like 2-4x slower depending on task
Show more
Gemini 4 Carbon Voxel Pagoda
Carbon has been in arena since yesterday, so I will show off some outputs I got over today
It's a noticeable improvement over Argon but is incredibly slow, like 2-4x slower depending on task
Show more
T3 code devs are now ripping Fable 5.5 apparently 😂
This could just be an error or a typo or something, but partners do have access now so I wouldn't be surprised
In other words, Fable 5.5 will read your mind
Anthropic is preparing a new AI model that can better infer what users are trying to accomplish, even when their instructions are vague.
The unreleased model also shows stronger reasoning and more creative problem-solving abilities, according to a person who has used it.
Full story:
Show more
Minimax M3.1 Flash scores 42 on the AA intelligence index
This doesn't look too impressive based off of AA's benchmarks, this is a bit of a letdown
Minimax M3.1 Flash benchmarks
This looks quite impressive tbf for a flash model
Minimax M3.1 Flash benchmarks
This looks quite impressive tbf for a flash model
Hello Minimax M3.1 flash (space bunny)
This has been in their docs for a long time now and is finally coming out of preview
It's apparently going to be a really nice jump up too so i'm expecting big things
Show more
🚨 The stealth model "Space Bunny" is officially a MiniMax model
MiniMax’s official "M3-a" deployment is returning the model: "space-bunny" across its benchmark runs
Given M3.1 is already staged for release in MiniMax Code, "M3-a" will be "M3-1"
This comes out of stealth testing Wednesday, i'd expect it to drop the same day
Show more
I actually wonder when GPT-6.1 Astra will drop
Anthropic are currently unchallenged and OpenAI have put a good amount of their compute into trying to fix their safety and alignment issues
I really hope they drop something big by the end of the month
Show more
I think we will see Fable 5.5 relatively soon
Not because Anthropic actually need to drop it (nothing comes close to Opus 5.5)
It will be purely for their IPO and to rub OpenAI's face in the fact that they're better currently
Show more
Had a lot of Anthropic employees tell me this is just goodwill
I still feel like they are hiding something internally to be this comfortable with letting money go, why do it now ?
What on earth do Anthropic have internally to be doing this
Anyone else noticing their Codex usage going unusually fast the past few days or is it just me ?
I think we will see Fable 5.5 relatively soon
Not because Anthropic actually need to drop it (nothing comes close to Opus 5.5)
It will be purely for their IPO and to rub OpenAI's face in the fact that they're better currently
Show more
What on earth do Anthropic have internally to be doing this
This is probably GPT Image 3 then
aurora-red-alpha is a faster model but worse quality
aurora-green-alpha is slower and better quality
Haven't managed to test blue out yet
All three of the new ChatGPT model are image gen models ??
aurora-blue-alpha
aurora-green-alpha
aurora-red-alpha
They are all testing in arena at the minute
