BestERP

MythoMax L2 13B

overall
0.6
(21)
nsfw
0.7
(14)
role playing
0.8
(14)
story writing
0.6
(10)

Overview

This model is proficient at both role-playing and story writing due to its unique nature.

Quantized versions

Submit your review

2.5 Not good

Pick which version of the model you base your review on.

Reviews

5.0

Using 6.13bpw exl2 quants, so far so good.

3.0

Best at prose but kinda old now. better alternatives out there.

0.0

It's old, and it's dumb. Most of his knowledge is about erp, and after a very short while it feels like the model is "wearing the mask of the characters" rather than "being the characters". There are a ton of better 13B models now, you're missing out if you still use MythoMax.

2.5

Oldie goldie, great at SFW roleplaying but really wants to take it slow as you want to engage in NSFW.

1.5
quant: other

Its old. It was one of the best models, but meanwhile are many better models out there. Its not human like and lacks in every category. I cant recommend it nowadays.

3.0

For its time, it wasn’t a bad model. It actually started to feel like we were taking these things for once and the character cards were being listened to. Overall, however, it’s relatively terse, and lacks detail compared to today’s models. Was good while it lasted.

2.5

So, what do you guys recommend now if this model is an oldie and goldie?

0.5
full precision

This is absolutely garbage. The only thing that its good for is maybe basic conversation. Otherwise, I don't recommend this model.

5.0

Look at the OpenRouter rankings. It has been number one since the beginning of this year, and has been number one so far. It's still March. The result is that many people still use it. Anyone can talk based on their preferences without looking at the numbers. It's actually an old model, but if its performance is poor, it should have been surpassed by a model in the same price range with similar maintenance costs. I know Claude, GPT, and Goliathe are great, but it's still a great model in the 13B genre.

4.0

Transformers 4 Bit quantization, works quite well...