Kimi K3.1 Pro on X
Moonshot AI · launched AnnouncementEventModel page
Launch vibes
The launch vibe for Kimi K3.1 Pro was split, and most of the posts were one-shot 3D or web builds run side by side with Claude Opus 5.5. People credited it with strong 3D work at a fraction of Opus's price, and some picked it outright. The doubts were about results and time: several said Opus still produced the better build, runs on max were slow or costly, and one camera fix failed after more than an hour. Of the 18 posts quoted here, 3 are positive, 10 mixed and 5 negative.
Several of these posts follow the same “same prompt, one shot, Kimi vs Opus” format, which may point to a coordinated campaign. None is marked as paid.
Pranav Reddy @AI_ScreeningPositive
“Kimi came out of nowhere and is now competing at the frontier.”
Nathan @nathangu76Mixed
“Kimi K3.1 Pro: 13m 02s · $1.26 … Opus 5.5: 19m 58s · $2.95 … I think Opus won clearly”
Michael Diestelberg @mdiestelbergPositive
“Kimi K3.1 Pro creates really nice landing pages and presentations.”
Yoseph Franco @arncofMixed
“Kimi K3.1 isn't the Astra killer. It's the Opus discount.”
whydeso @whydesoNegative
“11 minutes later: "Too many people are chatting with Kimi right now." Claude made it instead.”
Tony @EnvolDevNegative
“My Lighthouse test on max: $23.42 … Opus: $53. GPT-6 Astra: $42. Kimi lost to both.”
Dev Ed @developedbyedNegative
“It spent an hour and 20 minutes and it failed … This model is terrible I'm sorry”
Rithesh @Retired654321Mixed
“Actually Kimi K3.1 Pro is a good model. I tested it on my test suite and it did good. It is a little bit expensive”
Dromexa @dromexaPositive
“Time: Opus 65m / Kimi 20m … Cost: Opus $8.56 / Kimi $2.33 … My pick: Kimi.”
Kai @hqmankMixed
“The one-shot result was honestly impressive. … The one-shot took almost 20 minutes. GPT 6 Astra did the same task in just over 6.”
Niro @NirobswMixed
“Kimi was almost 3x faster and 4x cheaper. … Opus wins on detail and atmosphere. Kimi wins on price and speed.”
Lumeo Labs @LumeoLabMixed
“kimi k3.1 pro is half the price of opus 5.5. it also took 4x longer to build the same game.”
dab1rd @Daab1rDMixed
“opus still wins where it counts … but head to head it was a tie”
The Bugged Dev @thebuggeddevMixed
“honestly, I didn’t expect it to get this close to the original. … it feels little rough to me unless you're very specific to it”
Serik @sshayahmetNegative
“I prefer Terra’s result over both Kimi and DeepSeek in these tests, and the speed difference is hard to ignore.”
Hype Meter @tinycrewgoMixed
“Claude's looked better. Kimi costs half as much per token.”
AJ @ItsmeAjayKVNegative
“I tested Kimi-K3.1 with 3 or 4 examples, but i'm back to DeepSeek.”
Cuth @ItsCuthulhuMixed
“the density is what the money buys. … If the scene is the deliverable, it earns it. If you are iterating, the bill and the clock add up fast.”