TheDrummer/Artemis-31B-v1.2

Join our Discord! https://discord.gg/BeaverAI

More than 10000 members strong 💪 A hub for users and makers alike!

---

Thank you to everyone who subscribed through Patreon. Your support helps me chug along in this brave new world.

Drummer is open for new opportunities (I'm a Software Engineer). Contact me through any of these channels: https://linktr.ee/thelocaldrummer

---

Drummer proudly presents...

Artemis 31B v1.2 🌙

Usage

• Gemma 4 31B template, thinking or non-thinking.
• Sampler wrangling may be required (or be conservative)
• Gemma thinking is razor sharp for RP
• Crowdsourced Samplers
• Share your samplers!

Links

• Original: https://huggingface.co/TheDrummer/Artemis-31B-v1.2
• GGUF: https://huggingface.co/bartowski/TheDrummer_Artemis-31B-v1.2-GGUF

Reception

Artemis v1.2 landed as the strongest Artemis yet — and, for several testers, the strongest Gemma 4 tune they've used. The recurring themes:

• Writing that doesn't feel like G4. Immersive, creative prose with genuinely human-sounding dialogue; multiple testers made it their daily driver, and one got so absorbed in the character they forgot to test darker themes at all.
• Reasoning that actually works. The think block pulls the right details and the response stays faithful to it — no cross-card personality bleed at ~32k, scenario adherence holding at 20k, callbacks to small details from far earlier in a chat, and one clean 100k handoff.
• Tracking and consistency. Plot and detail tracking well ahead of previous versions, far fewer swipes and edits needed, and the old redraft loop is gone. Tool calls and general-purpose use remain intact.
• Class above its size. Repeatedly compared favourably to models several weight classes up.

who's jane doe?

"This is @Drummer's best so far. No competition."

"v1q might just be the most capable RP model out there rn."

"The only real competition is GLM 5.1 in some aspects which is several weight classes above and this model is still going head to head against it."

Sytan - No Kings

"This is a cut above the rest. Holy shit this doesn't even feel like G4 anymore"

"Whatever the fuck you did to Artemis V1.2, unbelievably good work man. Seriously, this model is amazing"

"Jesus, I am actually floored at how good Artemis 1.2 is"

"Genuinely what the fuck did you smoke when you made Artemis 1.2? This model is miles ahead of any other G4 tune I have used. Its actually insane"

armstrong64

"RP is solved now anyway. Artemis 1.2 is here."

"Ok this is actually crazy good. May be the smartest and most fun G4 tune I've tried."

"Much more immersive prose and captivating character-wise than most G4 tunes I've tried."

"Easily my new daily driver. The writing is so damn good."

"Everyone should try the new Artemis. It's insanely good, I'm kinda blown away by it. The level of intelligence, creativity, the excellent writing, it's something else. And I'm just running it at q3ks, neutral-ish samplers."

"the new Artemis is so good I didn't even get around to nsfw stuff at first cause the character was just too interesting to talk to lol."

"It's very concerning how good Artemis 1.2 is. Drummer needs to start pacing the frontier."

Silicas

"This thing's reasoning is legit."

"it's such a step up I'd call it more like a v2."

at ~32k: "more coherent and creative than v1o… its reasoning is correctly pulling relevant details from each card without mixing them up."

OJVD

"Writing is almost on par with Skyfall 4.2… but Artemis v1q keeps track of plot/details much better… With Artemis v1q I barely have to swipe or edit output… this has now become my go-to model for most of my rps."

veldrin

"this model has the best RP ive used in a long time"

at 20k: "it's still bringing up little things from earlier on in my conversations."

Cypaloma

"we really are dancing around glm5 quality around here… extremely high level model, seems to be working fine over here, no release blockers at all."

Rovagor

"Latest fine tune is so fun to use. Feels unique. Feels 'cinema'. It's just a blast to use so far!"

boop

"at 20k context it's still adhering to the scenario and representing characters consistently with no broken think. That feels better behaved than any other versions before to me."

N4M3L3S

"This model is peak, top quality scenery."

"if you can fit Artemis V1q 31B it is more than worth it, im using it at 5 t/s but it's the funniest"

nesymerp1

"v1q's dialouge is much more human like, compared to any Gemma4 tune. How the fuck did you do it."

Faust

"v1q is probably the most consistant so far out of what I've tested."

"your fine-tune still does tool calls fine, and generally works as general-purpose still too."

Amp

"Feels like the best artemis so far."

"listens to main prompt/ system prompt very well imo."

CinderGrace (mixed overall)

"the thinking is still top notch with this tune… I feel it's better than even stock gemma 4 31B"

"it opened up that prose massively, in the right direction."

Olede

"subjectively better than other Artemis versions. Decently creative, fun to talk with… better cards produce really interesting and deep RPs."

heftig

"I concur that this tune is brilliant."

Smir K. Shiverson

"Good character adherence, RP, smarts, reasoning."

Gojo Satoru

"Artemis v1q PEAK. 🔥"

Pjoak

"Artemis shiny work good want more"

config-v1q