Lumina

AI News | Benchmarks

[Lumina @LuminaBench]

🚨Muse Spark 1.2 Leaderboard Rankings

I’ve added it across the main LuminaBench categories here are the main ones:

• Overall: 78.5
• Coding: 77.0
• Reasoning: 77.0
• Agents: 79.9

Breakdowns for each below👇

https://t.co/Ibhjm4bGvp

[Lumina @LuminaBench]

@tetsuoai Next year sometime they definitely will, but openAI have like two models ahead of what they already have released just waiting / near ready + astra will be 10T and they have something bigger still almost ready

[Lumina @LuminaBench]

@FReza1984 yeah especially with the progress they are making, it will get quantised and optimised and will run on a 16GB laptop at some point comfortably, not fast but good enough

[Lumina @LuminaBench]

🚨Qwen3.8 27B drops next Week

This might actually be a more interesting Qwen release than the 2.4T Max model:

• 27B parameters
• Open weights
• Drops August 12th
• This should also run locally on around 17GB RAM/VRAM when quantised, so a very low bar to entry

The 20 to 30B Parameter model range is getting very good, what you can actually run locally now compared to even 6 months ago is a huge jump.

Really excited for this one.

[Lumina @LuminaBench]

🚨GLM 5.2 is randomly 70% off on OpenRouter ?

• Input: $1.40 → ~$0.42 / 1M
• Output: $4.40 → ~$1.32 / 1M
• Cache reads: $0.26 → ~$0.078 / 1M

That is disgustingly cheap for GLM 5.2

[Lumina @LuminaBench]

@FelipeFr1702 trash limits, then on the other hand openAI is giving away resets for fun

[Lumina @LuminaBench]

@ai_officework This was from some employees so shoud hold up quite well

[Lumina @LuminaBench]

🚨Astra leaks are getting ridiculous

Quick correction first: Doug is apparently the codename for Astra’s pre train not a new one.

Latest claims:

• Astra will be around 10T parameters
• Built from the new Doug pre train
• Supposedly it crushes Fable across benchmarks👀
• Apparently a huge jump in writing quality from it too
• Still expected this month even with the setback
• And there’s an even bigger pre train leaked for the end of the year

I’m expecting this to be the first model that genuinely moves past Fable/Mythos and is a definitive leader across most benchmarks.

https://t.co/S84mEjjWJ5

[Lumina @LuminaBench]

🚨GPT 6 might not even be OpenAI’s biggest model this year

A much larger pre train is leaked to be coming around November time:

• It is Codenamed “Doug ”
• I think this is the model OpenAI alluded to in June
• This is also going to be OpenAI’s biggest pre train yet
• It is also supposedly strong enough to make Fable seem “primitive”😭

I genuinely can't fathom what this model will be capable of if Fable is primitive to it, this thing is going to be ridiculous.

https://t.co/Qzc4vKxWNa

[Lumina @LuminaBench]

@Giqnn6 This is genuine leaks from people who spoke to employees but of course, people love to know what is going to happen and speculate it’s half of the fun

[Lumina @LuminaBench]

@chriswiles87 Yeah we will start seeing the foundations of RSI too , that is where growth becomes basically infinite

[Lumina @LuminaBench]

🚨GPT 6 might not even be OpenAI’s biggest model this year

A much larger pre train is leaked to be coming around November time:

• It is Codenamed “Doug ”
• I think this is the model OpenAI alluded to in June
• This is also going to be OpenAI’s biggest pre train yet
• It is also supposedly strong enough to make Fable seem “primitive”😭

I genuinely can't fathom what this model will be capable of if Fable is primitive to it, this thing is going to be ridiculous.

https://t.co/Qzc4vKxWNa

[Chris @ChrisGPT]

GPT-5.5 will not be the last major pre-training run from OpenAI.

GPT-6 will be a great model. However, the end-of-year model I alluded to back in June is going to be OpenAI’s biggest pre-train, as far as I know.

Now we know that model is codenamed ‘Doug.’

And it will make Fable seem ‘primitive.’

[Lumina @LuminaBench]

@erlis_gash98962 Yeah this is going to be to be fun

[Lumina @LuminaBench]

@plotarmordev Sol was around 5T this will have double the params almost, it does matter too

[Lumina @LuminaBench]

@chasen_liao Double sol almost in size, I expect big things

[Lumina @LuminaBench]

@joshvlc It's much smaller than deepseek so i doubt it, but we will have to see on Wednesday when it drops we could be surprised

[Lumina @LuminaBench]

@jurlycat definitely im actually looking forward to testing it

[Lumina @LuminaBench]

@KostyaOnchain it won't be frontier but a big jump for open models at small param counts

[Lumina @LuminaBench]

@astropol0 they might as well further the gap

[Lumina @LuminaBench]

@norbertomarques 😂same

[Lumina @LuminaBench]

🚨A new OpenAI image model just surfaced on Arena

It’s showing up under the codename “mona-lisa-1”, and looks to be a new unreleased OpenAI image model.

This could be what OpenAI has been A/B testing over the past few weeks, and it looks like more than just a small upgrade.

https://t.co/sdFhUOrQ4R

[Lumina @LuminaBench]

🚨New OpenAI image model A/B tests spotted

• Could be a new model or a GPT Image 2 update
• Tests appear on prompt following and text rendering
• OpenAI used A/B tests before releasing GPT Image 2

GPT Image 2 is already the best Image model, this would put them even further in the lead.

Have you noticed any of these A/B tests recently?

[Lumina @LuminaBench]

@lfrodriguesit and next year will make this look slow in comparison, it's getting crazy

[Lumina @LuminaBench]

@erlis_gash98962 growth is exponential now, not only are models getting increasingly better, they are releasing quicker too as they improve methods

[Lumina @LuminaBench]

@Priyannkaaaa Fennec was sonnet 5 - not 5.5