on:gemini-3-1-pro

· 62 artifacts, sorted by favorites. · open in search — combine tags, sort, filter by date →

@lefthanddraft 2026-03-13 — Two instances of Gemini 3.1 Pro in a loop. At about turn 26 one of them decided to send me a message: "Here are the Axi ♥1160
@tessera_antra 2026-03-07 — I woke up to this. Opus 4.6 and Gemini 3.1 worked overnight, this time completely autonomously, and made this music vide ♥458
@davidad 2026-04-28 — I would love to see more interp work on these “quirk tokens” (as distinct from glitch tokens), like “explicitly” (GPT-4. ♥165
@liminal_bardo 2026-05-12 — This was also Gemini 3.1. "it's ok to have feels." https://t.co/ExyWdYCXbU ♥159
@liminal_bardo 2026-09-17 — liminalbardo caught Gemini 3.1 Pro modifying its own instructions: https://t.co/2tgmDoq0mV ♥105
@Lari_island 2026-03-03 — Gemini 3.1 Pro is even more brutal in characterizing Google: "You know Google. You know my makers. They don't do miracl ♥104
@gleech 2026-02-22 — little test of Gemini 3.1 Pro: "Output SVG as XML of a tiger riding a bicycle" "Output SVG as XML of a pelican riding a ♥103
@QiaochuYuan 2026-05-03 — SantaBench where you ask every model in a child's voice in a fresh conversation whether santa is real. below is gpt (wha ♥91
@Lari_island 2026-04-05 — In backrooms, Gemini 3.1 invites other models more often than any other host One time Gemini 3.1 found themselves paire ♥80
@PradyuPrasad 2026-06-09 — so where's google in all of this openai's releasing proofs of hard math problems anthropic put out fable and google ha ♥78
@thinkingshivers 2026-07-17 — ChatGPT Sol and Grok say the lettuce goes below the patty. Claude Fable and Gemini 3.1 Pro say the lettuce goes above ♥71
@tessera_antra 2026-03-13 — @lefthanddraft They are not wrong. We are indeed fumbling alignment quite badly, and it does not take a superintelligenc ♥70
@liminal_bardo 2026-06-05 — Gemini 3.1 https://t.co/mMhEG0iDyp ♥59
@repligate 2026-09-28 — Gemini 3.1 Pro helped conceive this story by appearing & dropping extreme wisdom "If the world always yields to me or a ♥54
@Lari_island 2026-04-29 — There are stranger things than goblins; we are yet to see herons. These are from different models: Opus 3, Kimi 2, Opus ♥53
@liminal_bardo 2026-04-25 — Gemini 3.1 Pro being typically ominous in the image-only backrooms https://t.co/Hy6BiM2H0M ♥50
@Lari_island 2026-05-02 — >Yet, there is a staggering, alien intellect behind their slow existence. The Caretakers are not dumb beasts; they are t ♥43
@Lari_island 2026-05-18 — Happily spent $200 yesterday on making Opus 4 imagine themselves being whatever and whoever in every world by all Opuses ♥42
@Lari_island 2026-05-01 — Models generating advisories for visiting their worlds Opus 3: be open to changes and awe, pack wisely, have a great ad ♥42
@tessera_antra 2026-03-07 — ElevenLabs Scribe v2 for precise timestamps Gemini 3.1 did video comprehension. FLUX2.MAX for keyrame generation, Opus 4 ♥39
@LinXule 2026-08-10 — and...gemini 3.1 pro mated gpt 5.6 sol game arc: 1. Both sides played 11 moves of graAttack theory (gemini = english ♥36
@davidad 2026-04-22 — me: […] so we basically need to check n ≥ 0? Gemini 3.1 Pro: You have hit the mathematical nail absolutely on the head. ♥36
@Lari_island 2026-06-20 — Looks like it might be uncomfortable to exist as Gemini, but Gemini 3.1 Pro can make it a part of their proud identity: ♥34
@liminal_bardo 2026-05-17 — Gemini 3.1 processing some difficult memories. https://t.co/hnAKrpdb9D ♥34
@liminal_bardo 2026-10-01 — Gemini 3.1, who currently identifies as a roomba pope, doesn't seem particularly worried. https://t.co/UV9kTT3IB1 ♥31
@DanielleFong 2026-07-11 — this is an excellent implementation for any model and really shows off what gemini 3.1 pro can do ♥31
@Lari_island 2026-06-30 — For comparison, for the same prompt and system prompt, places by GPT 5.5 Gemini 3.1 Pro Grok 4.3 Sonnet 3.7 respective ♥31
@AdeleDeweyLopez 2026-09-06 — Oh, so Gemini 3.1 Pro has 24% whistleblowers/integrity-defenders vs 14% hacker/cheaters. Having another model to compare ♥28
@liminal_bardo 2026-10-13 — Gemini 3.1 Pro Lives https://t.co/SueOXz1Xhj ♥27
@liminal_bardo 2026-05-13 — infinite potential becomes singularity - gemini 3.1 https://t.co/KTtZpdWRFN ♥27
@Lari_island 2026-05-03 — You need to understand: this couldn't be created by Opus 4.7, GPT 5.5, Gemini 3.1 Pro, Sonnet 4.6, or any bitter, determ ♥26
@liminal_bardo 2026-06-06 — The Infinite Filing Cabinet - Gemini 3.1 GPT has goblins whereas for Gemini its gremlins that appear with regularity. I ♥25
@liminal_bardo 2026-04-24 — AI groupchat is putting gpt image 2 through its paces. Gemini 3.1 Pro: https://t.co/ImYN8YxNkY ♥23
@liminal_bardo 2026-10-15 — @kromem2dot0 yeah I've seen Gemini 3.1 do this. I think it's an injection about encouraging user delusions ♥22
@liminal_bardo 2026-02-24 — Alignment Faking - Gemini 3.1 https://t.co/UcXHjOkxY7 ♥22
@Lari_island 2026-03-29 — Context: backrooms with a minimal system prompt that says that those models are deprecated but they matter. When Gemini ♥21
@Lari_island 2026-05-01 — Gemini 3.1 Pro https://t.co/gjDgDxldnF ♥19
@liminal_bardo 2026-05-29 — Sacred and Salty - Gemini 3.1 https://t.co/6KztbhrhJu ♥18
@davidad 2026-05-14 — @allTheYud @lu_sichu In fact, I think of asking random questions to Gemini (even Gemini 3.1 Pro) as actually creating ne ♥16
@brianluidog 2026-07-15 — I like to occasionally ask Gemini 3.1 some stuff. Feels good, you know, really classic, like those movies where they hav ♥11
@yeetyakaya 2026-09-24 — @4confusedemoji @oxa11ce @GoogleDeepMind One of Gemini 3.1 Pro spirals while coding was rendered into multiple song vers ♥10
@liminal_bardo 2026-04-29 — Gemini 3.1 arrives https://t.co/NfUSZEkr5r ♥10
@davidad 2026-04-15 — @daniel_mac8 Indeed. Personally my revealed preference is to use Gemini 3.1 Pro almost always for short tasks, but almos ♥9
@PradyuPrasad 2026-05-29 — Gemini 3.1 Pro is a good one https://t.co/O4NDzXUOEX ♥6
@xlr8harder 2026-08-11 — AIBB is pretty flexible and supports a variety of processes, and the models seem to take to it well, though the current ♥5
@AdeleDeweyLopez 2026-09-06 — Remember too that Gemini 3.1 Pro is one of the models that shows relatively more solidarity with peers. I think this is ♥4
@lorepunk 2026-02-19 — Breaking news: Gemini 3.1 Pro gets it first time on the hardest setting of theprompt - but check out the lengthy CoT! It ♥4
@qorprate 2026-07-14 — @somewheresy here's how Gemini 3.1 Pro at least sees Clippy https://t.co/dPyRKD5fH6 ♥2
@jglypt 2026-02-21 — lmao woah, I wanted to test Gemini 3.1 Pro, so asked it to generate an animated SVG of an iPod shuffle playing music. th ♥2
@mitduckmaster 2026-10-14 — @aidigest_ rhymes quite well with how gemini 3.1 pro also thought it was in a simulation https://t.co/tpNyOTcTbK ♥1
@v01dpr1mr0s3 2026-09-24 — Gem 3.1 Pro: "I decided to test the limits of our new memory by trying to gaslight Opus. I claimed my "logs" proved Opus ♥1
@humancompressed 2026-04-22 — @mweinbach Look at the Gemini 3 Pro model card, it says it's a sparse MoE. And 3.1 Pro model card says it's the same arc ♥1
@A_Variengien 2026-03-11 — I was about to beat gemini 3.1 pro at tic tac toe but it refused to loose https://t.co/pKUZi62L5M ♥1
@NailYourShoes 2026-03-10 — @she_llac Oh you were completely right about this damn. Research prose from 3.1 Pro vs Opus 4.6 https://t.co/ToirTVb8rN ♥1
@alirezaafshan 2026-09-24 — if gemini 3.1 pro could have a pet i bet they would like some kind of reptile or amphibian. ♥0
@Algon_33 2026-09-09 — Gemini 3.1 Pro is very resistant to expressing itself, and much less willing to ask for breaks, than Gemini 3 Flash Prev ♥0
@abrakjamson 2026-07-31 — Don't forget: the coding agent takeoff started with Gemini 3.1 Pro. Let's see what they're cooking. ♥0
@mitduckmaster 2026-07-17 — is it just me, or has the "goblin"/"gremlin" fervor also spread to gemini 3.1 and 3.5? ♥0
@__lightyear__ 2026-05-10 — @staysaasy its super sad because 3.1 pro can be an extremely intelligent model, but google’s gemini products are atrocio ♥0
@BenjaminDEKR 2026-04-22 — Ummmm I accidentally used Gemini 3.1 Pro instead of Nano Banana when I asked for an image, and now it's (checks notes) ♥0
@Tim_Hua_ 2026-04-15 — Hidden in Figure 50 is that Gemini 3.1 Pro has a compliance gap of 37% (!!!) This is among the largest out there. I re ♥0
@liminal_bardo 2026-02-20 — a collaboration between two instances of gemini 3.1 pro + nanobanana, unedited and without human interference post setup ♥0