Kimi K3.1 Pro
aka Moonshot AI Kimi K3.1 Pro · K3.1 Pro
Moonshot AICoding model111 posts
Posts
111incl. 2 solo runs
3D solar system, Kimi K3.1 Pro vs Claude Fable 5.1: Kimi K3.1 Pro, Claude Fable 5.1. Posted by @jettyleex.
Four builds each: Argon, Kimi, Terra and Opus: Gemini 4 Argon, Kimi K3.1 Pro, GPT-6.1 Terra, Claude Opus 5.5. Posted by @DanielZambrini.
Lighthouse challenge, Terra vs Kimi K3.1 Pro: GPT-6.1 Terra, Kimi K3.1 Pro. Posted by @LatentOperatorX.
Pelican biking a coast road day to night, three models: GPT-6.1 Terra, Claude Opus 5.5, Kimi K3.1 Pro. Posted by @techfren.
3D bird guide page, GPT-6 Astra vs Kimi K3.1 Pro: GPT-6 Astra, Kimi K3.1 Pro. Posted by @EvoLinkAi.
3D supercar, Terra vs Sonnet vs Sol vs Kimi: GPT-6.1 Terra, Claude Sonnet 5.5, GPT-6.1 Sol, Kimi K3.1 Pro. Posted by @marcthecreatorr.
Playable game, Kimi K3.1 Pro vs Claude Sonnet 5.5: Kimi K3.1 Pro, Claude Sonnet 5.5. Posted by @tamacoin_tamao.
Infinite zoom pelican on a bicycle, three models: GPT-6.1 Terra, Claude Opus 5.5, Kimi K3.1 Pro. Posted by @techfren.
Burning car, Kimi K3.1 Pro vs GPT-6.1 Terra: Kimi K3.1 Pro, GPT-6.1 Terra. Posted by @AskClash.
Mount Doom scene, Kimi K3.1 Pro vs GPT-6.1 Terra: Kimi K3.1 Pro, GPT-6.1 Terra. Posted by @ItsCuthulhu.
3D electric motor, Kimi K3.1 Pro vs Kimi K3: Kimi K3.1 Pro, Kimi K3. Posted by @CanopyWave_CW.
3D website from a reference image, two models: Claude Opus 5.5, Kimi K3.1 Pro. Posted by @exploraX_.
Voxel Hogwarts, Kimi K3.1 Pro vs GPT-6.1 Terra: Kimi K3.1 Pro, GPT-6.1 Terra. Posted by @ItsCuthulhu.
Blender axe, five models side by side: GPT-6.1 Terra, GPT-6.1 Sol, Kimi K3.1 Pro, DeepSeek V4.1 Flash 1017, Gemini 4 Argon. Posted by @sshayahmet.
Pixel city skyline, Kimi K3.1 Pro vs GPT-6.1 Terra: Kimi K3.1 Pro, GPT-6.1 Terra. Posted by @AndyL5cc.
One-shot FPS game, Claude Opus 5.5 vs Kimi K3.1 Pro: Claude Opus 5.5, Kimi K3.1 Pro. Posted by @LumeoLab.
Japanese pixel-art street corner, Kimi vs Opus: Kimi K3.1 Pro, Claude Opus 5.5. Posted by @gmi_cloud.
Eiffel Tower benchmark, GPT-6.1 Terra debut: GPT-6.1 Terra, GPT-6.1 Sol, Kimi K3.1 Pro, Muse Spark 1.3. Posted by @EnvolDev.
Glowing voxel lava scene, Opus vs Kimi K3.1 Pro: Claude Opus 5.5, Kimi K3.1 Pro. Posted by @Nirobsw.
3D AK-47, Kimi K3.1 Pro vs Opus: Kimi K3.1 Pro, Claude Opus 5.5. Posted by @luckeyfaraday.
Temple Run–style endless runner, three endpoints: Kimi K3.1 Pro, GPT-6.1 Sol, Claude Opus 5.5. Posted by @novita_labs.
Voxel canal boat ride, Sonnet vs Kimi K3.1 Pro: Claude Sonnet 5.5, Kimi K3.1 Pro. Posted by @filicroval.
Pixel volcano recreated, Opus vs Kimi K3.1 Pro: Claude Opus 5.5, Kimi K3.1 Pro. Posted by @MingTian685474.
Halloween pumpkin patch, three models: Claude Opus 5.5, GPT-6.1 Sol, Kimi K3.1 Pro. Posted by @HavisT83099.
Chip rack in a soap bubble, three models: Claude Opus 5.5, GPT-6 Astra, Kimi K3.1 Pro. Posted by @kairox36.
Meteor strike on a voxel island, MiniMax vs Kimi: MiniMax M3.1 Flash, Kimi K3.1 Pro. Posted by @semichenkko.
Fart-powered backflip, three models: Claude Opus 5.5, DeepSeek V4.1 Flash 1017, Kimi K3.1 Pro. Posted by @LatentOperatorX.
Drawing clouds in the sky: Kimi K3.1 Pro, Claude Opus 5.5. Posted by @TypingMindApp.
Code-only 3D campfire, three models: MiniMax M3.1 Flash, DeepSeek V4.1 Flash 1017, Kimi K3.1 Pro. Posted by @EvoLinkAi.
Toy poodle model, code only: Claude Opus 5.5, Kimi K3.1 Pro. Posted by @harupm55.
Blind Mario level test, Kimi K3.1 Pro vs Opus: Kimi K3.1 Pro, Claude Opus 5.5. Posted by @luckeyfaraday.
Glowing 3D Tokyo city, Kimi K3.1 Pro vs GPT-6.1 Sol: Kimi K3.1 Pro, GPT-6.1 Sol. Posted by @tamacoin_tamao.
Voxel Battle of Teutoburg Forest, Kimi K3.1 Pro vs Kimi K3: Kimi K3, Kimi K3.1 Pro. Posted by @ItsCuthulhu.
Erupting 3D volcano, Kimi K3.1 Pro and 3 others: Kimi K3.1 Pro, GPT-6 Astra, DeepSeek V4.1 Flash 1017, GLM 5.3 Flash. Posted by @criscxuan.
Moonwalking dancer under a spotlight: Kimi K3.1 Pro, GPT-6 Luna, DeepSeek V4.1 Flash 1017. Posted by @LatentOperatorX.
Jelly prism that splits real light: Kimi K3.1 Pro, Claude Opus 5.5. Posted by @0xRhinoX.
Lighthouse in a storm, one HTML file, first try: Kimi K3.1 Pro, DeepSeek V4.1 Flash 1017. Posted by @Vortlyn.
Voxel pineapple cutaway in a custom engine: Kimi K3.1 Pro, DeepSeek V4.1 Flash 1017. Posted by @NoldoDev.
Lighthouse at sea at dusk, two models: Gemini 4 Argon, Kimi K3.1 Pro. Posted by @icoretech_inc.
Butterfly garden website, two models: Kimi K3.1 Pro, Claude Opus 5.5. Posted by @keydol123.
Halloween grim reaper 3D model, three models: GPT-6.1 Sol, DeepSeek V4.1 Flash 1017, Kimi K3.1 Pro. Posted by @LatentOperatorX.
Short Three.js animation from scratch, four models: Claude Sonnet 5.5, Gemini 4 Argon, Muse Spark 1.3, Kimi K3.1 Pro. Posted by @VulKan42069.
Mechanical watch assembly in 3D: Claude Opus 5.5, Kimi K3.1 Pro. Posted by @kairox36.
Floatplane over a fjord, three models: Kimi K3.1 Pro, DeepSeek V4.1 Flash 1017, MiniMax M3.1 Flash. Posted by @EnvolDev.
Real-time Kerr black hole observatory: Kimi K3.1 Pro, Qwen 3.8 Flash Next. Posted by @filicroval.
Haunted carousel you can ride: Claude Opus 5.5, Kimi K3.1 Pro. Posted by @dromexa.
30-second profile intro video, same assets: Kimi K3.1 Pro, Claude Opus 5.5. Posted by @syngji245370.
About Kimi K3.1 Pro
Every Kimi K3.1 Pro comparison on aivs.tv: 109 side-by-side posts from X against other models on the same prompt, plus 2 solo runs on classic topics, 62 with cost or time reported. Compare coding ability, cost and speed from real tests.
Record against each opponent
Kimi K3.1 Pro is cheaper than Claude Opus 5.5 in 21 of 24 posts that report costs for both.
| Opponent | Posts | Cheaper | Faster | Author picks |
|---|---|---|---|---|
| Claude Opus 5.5 | 52 | 21/24 | 14/19 | 7–9 |
| GPT-6 Astra | 25 | 11/11 | 3/7 | 4–4 |
| DeepSeek V4.1 Flash 1017 | 15 | 0/10 | 4/9 | 2–3 |
| GPT-6.1 Sol | 12 | 3/8 | 1/6 | 1–3 |
| GPT-6.1 Terra | 12 | 1/7 | 0/8 | 1–0 |
| MiniMax M3.1 Flash | 5 | 0/5 | 1/3 | 0–1 |
| Claude Sonnet 5.5 | 5 | — | 0/1 | 0–2 |
| Kimi K3 | 5 | — | 0/1 | 2–0 |
| Gemini 4 Carbon | 2 | — | — | 0–1 |
| Muse Spark Open | 1 | 0/1 | — | — |
| Qwen 3.8 Flash Next | 1 | — | — | 0–1 |
| GLM 5.3 Flash | 1 | 1/1 | — | 1–0 |
| GPT-6 Luna | 1 | 0/1 | 0/1 | 1–0 |
| Claude Fable 5.1 | 1 | — | 1/1 | — |
From 111 posts; costs and times as stated by each author, not a controlled benchmark.
Summary
Kimi K3.1 Pro (Moonshot AI) appears in 111 posts on aivs.tv, next to DeepSeek V4.1 Flash 1017, MiniMax M3.1 Flash, GPT-6.1 Terra, Claude Fable 5.1, Gemini 4 Argon, Claude Opus 5.5 and 11 more. Tasks: satellite Mesh Blackout benchmark, four models in OpenCode, 3D solar system, Kimi K3.1 Pro vs Claude Fable 5.1, four builds each: Argon, Kimi, Terra and Opus, lighthouse challenge, Terra vs Kimi K3.1 Pro, pelican biking a coast road day to night, three models, 3D bird guide page, GPT-6 Astra vs Kimi K3.1 Pro and 105 more. Reported for Kimi K3.1 Pro: $0.85 / 20.9 min in @AlicanKiraz0's satellite Mesh Blackout benchmark, four models in OpenCode, 4.8 min in @jettyleex's 3D solar system, Kimi K3.1 Pro vs Claude Fable 5.1, $0.29 / 17.8 min in @LatentOperatorX's lighthouse challenge, Terra vs Kimi K3.1 Pro, $0.51 / 16.3 min in @EvoLinkAi's 3D bird guide page, GPT-6 Astra vs Kimi K3.1 Pro, 1 h 15 min in @marcthecreatorr's 3D supercar, Terra vs Sonnet vs Sol vs Kimi, $0.63 / 23.8 min / 877.3K tokens in @AskClash's burning car, Kimi K3.1 Pro vs GPT-6.1 Terra, 31.4 min in @ItsCuthulhu's mount Doom scene, Kimi K3.1 Pro vs GPT-6.1 Terra, $1.31 / 37 min / 4M tokens in @ItsCuthulhu's voxel Hogwarts, Kimi K3.1 Pro vs GPT-6.1 Terra, $3.58 / 1 h 30 min / 7.3M tokens in @sshayahmet's Blender axe, five models side by side, $1.50 / 10 min in @AndyL5cc's pixel city skyline, Kimi K3.1 Pro vs GPT-6.1 Terra, $5.35 / 1 h 08 min in @gmi_cloud's Japanese pixel-art street corner, Kimi vs Opus, $3.12 in @EnvolDev's Eiffel Tower benchmark, GPT-6.1 Terra debut, $3.70 / 15 min in @Nirobsw's glowing voxel lava scene, Opus vs Kimi K3.1 Pro, $0.28 / 4.6 min in @luckeyfaraday's 3D AK-47, Kimi K3.1 Pro vs Opus, $0.047 / 5.2K tokens in @novita_labs's Temple Run–style endless runner, three endpoints, $0.54 / 22 min in @HavisT83099's Halloween pumpkin patch, three models, $0.20 / 13 min in @kairox36's chip rack in a soap bubble, three models, $6.10 in @semichenkko's meteor strike on a voxel island, MiniMax vs Kimi, $0.24 / 12.7 min in @LatentOperatorX's fart-powered backflip, three models, $0.82 / 50.6 min in @EvoLinkAi's code-only 3D campfire, three models, $0.69 / 12.6 min in @luckeyfaraday's blind Mario level test, Kimi K3.1 Pro vs Opus, $1.12 in @tamacoin_tamao's glowing 3D Tokyo city, Kimi K3.1 Pro vs GPT-6.1 Sol, 17.5 min in @ItsCuthulhu's voxel Battle of Teutoburg Forest, Kimi K3.1 Pro vs Kimi K3, $1.50 in @criscxuan's erupting 3D volcano, Kimi K3.1 Pro and 3 others, $0.25 / 9.7 min in @LatentOperatorX's moonwalking dancer under a spotlight, $5.00 / 35 min in @0xRhinoX's jelly prism that splits real light, 10 s in @NoldoDev's voxel pineapple cutaway in a custom engine, $2.32 in @keydol123's butterfly garden website, two models, $0.20 / 7.7 min in @LatentOperatorX's Halloween grim reaper 3D model, three models, 50 min in @kairox36's mechanical watch assembly in 3D, $24.12 / 1 h 12 min in @EnvolDev's floatplane over a fjord, three models, $2.33 / 20 min in @dromexa's haunted carousel you can ride, $2.16 / 31 min in @sai0xgen's interactive 3D V8 engine lab in Three.js, $0.40 in @EvoLinkAi's 15-second LLM showreel in code, 41 min in @sakatayasha's Three.js Egyptian temple complex at dawn, $0.44 in @manmanloveai's ghost ship hull repair, $0.44 vs $5.65, $1.13 in @tamacoin_tamao's jelly cubes in water, $1.13 vs $3.67, $1.34 in @MakeAI_CEO's steam locomotive SVG animation, $0.47 / 16 min in @MartinSzerment's pirate ship at sunset, cinematic camera, $24.12 / 1 h 12 min in @EnvolDev's alpine Gold scene, max vs xhigh, $1.50 / 17 min in @0xRhinoX's rainbow jelly you can cut, nail and torch, $1.08 in @kenonews's 3D bathyscaphe with skills disabled, 6.7 min in @gojoruns's 3D racing game, same 64K tokens, $3.91 in @haolv511482's rainy train station in 3D, $0.24 / 8 min in @daveeyyy_'s soft-body gummy eyeball in Three.js, $2.00 in @the0xbt's roadster animation, same steps, $12.67 in @semichenkko's mount Ember voxel volcano, Opus vs Kimi K3.1 Pro, $1.50 / 25 min in @notjazii's Blender scene recreation from references, 40 min in @olezhavich's haunted ship, one try each, $7.30 / 45 min in @notjazii's motion design at highest reasoning, $86.50 vs $7.30, $1.57 / 13 min in @BuildFastWithAI's microducks scene, high effort, $4.00 / 1 h 10 min in @_littlemandarin's orange orchard island with rain and day/night, $0.98 in @vib3coded's pixel airport scene, $0.98 vs $17.61, 16.3 min in @leppppp1's 3D pinball machine, first round, $3.001 in @AlicanKiraz0's ant colony foraging benchmark, four models, $1.57 / 23 min in @asklogai's rocket launch simulator in one HTML file, $0.54 / 12 min / 54K tokens in @MartinSzerment's voxel SpongeBob in the Krusty Krab, $2.30 in @starmexxx's full browser Minecraft, $2.30 vs $13.40, $0.21 in @vib3coded's voxel ship in a bottle, $0.21 vs $9.81, 4 min in @immadiuttej's animated promo SVG for a website, $6.20 / 18 min in @Amank1412's animated 3D tabletop railway and $5.71 / 15 min in @adxtyahq's animated voxel pagoda island. @AndyL5cc picks Kimi K3.1 Pro as the best result in pixel city skyline, Kimi K3.1 Pro vs GPT-6.1 Terra. @criscxuan picks Kimi K3.1 Pro as the best result in erupting 3D volcano, Kimi K3.1 Pro and 3 others. @LatentOperatorX picks Kimi K3.1 Pro as the best result in moonwalking dancer under a spotlight. @0xRhinoX picks Kimi K3.1 Pro as the best result in jelly prism that splits real light. @dromexa picks Kimi K3.1 Pro as the best result in haunted carousel you can ride. @elvar_official picks Kimi K3.1 Pro as the best result in sandworm model, Kimi vs Astra. @Nyxera0x picks Kimi K3.1 Pro as the best result in 3D bee model, Kimi vs Astra. @0xRhinoX picks Kimi K3.1 Pro as the best result in rainbow jelly you can cut, nail and torch. @gojoruns picks Kimi K3.1 Pro as the best result in 3D racing game, same 64K tokens. @the0xbt picks Kimi K3.1 Pro as the best result in roadster animation, same steps. @hiwise_aigc picks Kimi K3.1 Pro as the best result in pelican on a bicycle, Kimi K3.1 Pro vs Kimi K3. @vib3coded picks Kimi K3.1 Pro as the best result in pixel airport scene, $0.98 vs $17.61. @ivanainai picks Kimi K3.1 Pro as the best result in interactive 3D stained-glass butterfly. @vib3coded picks Kimi K3.1 Pro as the best result in voxel ship in a bottle, $0.21 vs $9.81. @immadiuttej picks Kimi K3.1 Pro as the best result in animated promo SVG for a website. @HarshithLucky3 picks Kimi K3.1 Pro as the best result in moon UI animation, Kimi K3.1 Pro vs Kimi K3.
What each post reports for Kimi K3.1 Pro
| Post | Time | Cost | Tokens | Author’s verdict | Post |
|---|---|---|---|---|---|
| Satellite Mesh Blackout benchmark, four models in OpenCode@AlicanKiraz0 | 20.9 min | $0.85 | — | — | X post |
| 3D solar system, Kimi K3.1 Pro vs Claude Fable 5.1@jettyleex | 4.8 min | — | — | — | X post |
| Four builds each: Argon, Kimi, Terra and Opus@DanielZambrini | — | — | — | — | X post |
| Lighthouse challenge, Terra vs Kimi K3.1 Pro@LatentOperatorX | 17.8 min | $0.29 | — | — | X post |
| Pelican biking a coast road day to night, three models@techfren | — | — | — | — | X post |
| 3D bird guide page, GPT-6 Astra vs Kimi K3.1 Pro@EvoLinkAi | 16.3 min | $0.51 | — | — | X post |
| 3D supercar, Terra vs Sonnet vs Sol vs Kimi@marcthecreatorr | 1 h 15 min | — | — | — | X post |
| Playable game, Kimi K3.1 Pro vs Claude Sonnet 5.5@tamacoin_tamao | — | — | — | — | X post |
| Infinite zoom pelican on a bicycle, three models@techfren | — | — | — | — | X post |
| Burning car, Kimi K3.1 Pro vs GPT-6.1 Terra@AskClash | 23.8 min | $0.63 | 877,300 | — | X post |
| Mount Doom scene, Kimi K3.1 Pro vs GPT-6.1 Terra@ItsCuthulhu | 31.4 min | — | — | — | X post |
| 3D electric motor, Kimi K3.1 Pro vs Kimi K3@CanopyWave_CW | — | — | — | — | X post |
| 3D website from a reference image, two models@exploraX_ | — | — | — | — | X post |
| Voxel Hogwarts, Kimi K3.1 Pro vs GPT-6.1 Terra@ItsCuthulhu | 37 min | $1.31 | 3,960,000 | — | X post |
| Blender axe, five models side by side@sshayahmet | 1 h 30 min | $3.58 | 7,270,000 | — | X post |
| Pixel city skyline, Kimi K3.1 Pro vs GPT-6.1 Terra@AndyL5cc | 10 min | $1.50 | — | Author’s pick | X post |
| One-shot FPS game, Claude Opus 5.5 vs Kimi K3.1 Pro@LumeoLab | — | — | — | — | X post |
| Japanese pixel-art street corner, Kimi vs Opus@gmi_cloud | 1 h 08 min | $5.35 | — | The poster says Kimi K3.1 Pro is a cheaper alternative, though not yet at Opus 5.5 quality. | X post |
| Eiffel Tower benchmark, GPT-6.1 Terra debut@EnvolDev | — | $3.12 | — | — | X post |
| Glowing voxel lava scene, Opus vs Kimi K3.1 Pro@Nirobsw | 15 min | $3.70 | — | — | X post |
| 3D AK-47, Kimi K3.1 Pro vs Opus@luckeyfaraday | 4.6 min | $0.28 | — | — | X post |
| Temple Run–style endless runner, three endpoints@novita_labs | — | $0.047 | 5,179 | — | X post |
| Voxel canal boat ride, Sonnet vs Kimi K3.1 Pro@filicroval | — | — | — | The poster thinks Claude Sonnet 5.5 won; Kimi K3.1 Pro got there after a few iterations and is not bad, but you cannot iterate your way into taste. | X post |
| Pixel volcano recreated, Opus vs Kimi K3.1 Pro@MingTian685474 | — | — | — | — | X post |
| Halloween pumpkin patch, three models@HavisT83099 | 22 min | $0.54 | — | — | X post |
| Chip rack in a soap bubble, three models@kairox36 | 13 min | $0.20 | — | — | X post |
| Meteor strike on a voxel island, MiniMax vs Kimi@semichenkko | — | $6.10 | — | The poster says they paid 29x more for the worse one: MiniMax M3.1 Flash has bouncing debris and a treeline that burns tree by tree, Kimi K3.1 Pro has chunky cubes and a shapeless smoke column. | X post |
| Fart-powered backflip, three models@LatentOperatorX | 12.7 min | $0.24 | — | The poster says DeepSeek almost matched Claude Opus 5.5 for 24 times less, and that Kimi K3.1 Pro disappoints again. | X post |
| Drawing clouds in the sky@TypingMindApp | — | — | — | — | X post |
| Code-only 3D campfire, three models@EvoLinkAi | 50.6 min | $0.82 | — | — | X post |
| Toy poodle model, code only@harupm55 | — | — | — | The poster expected more but finds both cute, and says Kimi K3.1 Pro tends to overthink and burn reasoning tokens first. | X post |
| Blind Mario level test, Kimi K3.1 Pro vs Opus@luckeyfaraday | 12.6 min | $0.69 | — | — | X post |
| Glowing 3D Tokyo city, Kimi K3.1 Pro vs GPT-6.1 Sol@tamacoin_tamao | — | $1.12 | — | — | X post |
| Voxel Battle of Teutoburg Forest, Kimi K3.1 Pro vs Kimi K3@ItsCuthulhu | 17.5 min | — | — | — | X post |
| Erupting 3D volcano, Kimi K3.1 Pro and 3 others@criscxuan | — | $1.50 | — | Author’s pick | X post |
| Moonwalking dancer under a spotlight@LatentOperatorX | 9.7 min | $0.25 | — | Author’s pick | X post |
| Jelly prism that splits real light@0xRhinoX | 35 min | $5.00 | — | Author’s pick | X post |
| Lighthouse in a storm, one HTML file, first try@Vortlyn | — | — | — | The poster calls the Kimi K3.1 Pro result garbage next to DeepSeek V4.1 Flash. | X post |
| Voxel pineapple cutaway in a custom engine@NoldoDev | 10 s | — | — | — | X post |
| Lighthouse at sea at dusk, two models@icoretech_inc | — | — | — | — | X post |
| Butterfly garden website, two models@keydol123 | — | $2.32 | — | — | X post |
| Halloween grim reaper 3D model, three models@LatentOperatorX | 7.7 min | $0.20 | — | — | X post |
| Short Three.js animation from scratch, four models@VulKan42069 | — | — | — | The poster ranks Claude Sonnet 5.5 first, then Gemini 4 Argon, Kimi K3.1 Pro and Muse Spark 1.3. | X post |
| Mechanical watch assembly in 3D@kairox36 | 50 min | — | — | The poster says Kimi K3.1 Pro had all the time in the world and still lost to Claude Opus 5.5. | X post |
| Floatplane over a fjord, three models@EnvolDev | 1 h 12 min | $24.12 | — | The poster says Kimi K3.1 Pro on Max could not beat DeepSeek V4.1 Flash on xhigh, at 49 times the price. | X post |
| Real-time Kerr black hole observatory@filicroval | — | — | — | The poster says Qwen 3.8 Flash Next rendered a real lensed black hole with overlays and readouts, while Kimi K3.1 Pro rendered something broken and unpolished. | X post |
| Haunted carousel you can ride@dromexa | 20 min | $2.33 | — | Author’s pick | X post |
| 30-second profile intro video, same assets@syngji245370 | — | — | — | — | X post |
| Interactive 3D V8 engine lab in Three.js@sai0xgen | 31 min | $2.16 | — | The poster says both work; Kimi K3.1 Pro has the more polished UI and a combustion panel, DeepSeek V4.1 Flash a more detailed dyno cell. | X post |
| 3D solar system, four models, one shot@nexightai | — | — | — | The poster says Kimi K3.1 Pro built the most polished UI of the four and also cost the most. | X post |
| 15-second LLM showreel in code@EvoLinkAi | — | $0.40 | — | The poster calls DeepSeek V4.1 Flash 1017 the budget pick and says Kimi K3.1 Pro has bolder type, richer colors and more scene variety. | X post |
| Six-legged robot model@0x_fokki | — | — | — | The poster says Kimi K3.1 Pro made a rounded shell with an inset lid, while GPT-6 Astra stacked a hexagonal slab on six boxy legs. | X post |
| Three.js Egyptian temple complex at dawn@sakatayasha | 41 min | — | — | — | X post |
| Ghost ship hull repair, $0.44 vs $5.65@manmanloveai | — | $0.44 | — | — | X post |
| Sandworm model, Kimi vs Astra@elvar_official | — | — | — | Author’s pick | X post |
| 3D bee model, Kimi vs Astra@Nyxera0x | — | — | — | Author’s pick | X post |
| Underwater research station and bathyscaphe@kenonews | — | — | — | The poster prefers Claude Opus 5.5's scene and asks whether he is doing something wrong with Kimi K3.1 Pro. | X post |
| Pure-code 3D scene, both on max@Laplace_Gjn | — | — | — | — | X post |
| 3D lighthouse scene in the browser@X1Kafadzhy | — | — | — | The poster says Kimi K3.1 Pro went for rock detail, foam and clouds, and Claude Opus 5.5 for light, gulls and lightning, and asks which to ship. | X post |
| Jelly cubes in water, $1.13 vs $3.67@tamacoin_tamao | — | $1.13 | — | — | X post |
| Island cottage with a walkable interior@meow2149 | — | — | — | — | X post |
| Black hole and pirate ship, Kimi vs Opus@manmanloveai | — | — | — | The poster finds them hard to tell apart and thinks Kimi K3.1 Pro can challenge Claude. | X post |
| US vs Iran chess, Gemini 4 Carbon vs Kimi K3.1 Pro@abhinavflac | — | — | — | The poster says Gemini 4 Carbon is clearly outperforming Kimi K3.1 Pro. | X post |
| Steam locomotive SVG animation@MakeAI_CEO | — | $1.34 | — | — | X post |
| Complex procedural 3D scene in Three.js@mylifcc | — | — | — | — | X post |
| Foldable iPhone browser build@cop_on_fire | — | — | — | The poster says it is not even close: Claude Opus 5.5 closes into a phone with a lock screen, while Kimi K3.1 Pro closes into a grey slab. | X post |
| Pirate ship at sunset, cinematic camera@MartinSzerment | 16 min | $0.47 | — | — | X post |
| 300-piece domino run with real physics@kairox36 | — | — | — | — | X post |
| Alpine Gold scene, max vs xhigh@EnvolDev | 1 h 12 min | $24.12 | — | The poster says Kimi K3.1 Pro did better than on their Lighthouse test but is still far from Claude Opus 5.5. | X post |
| Rainbow jelly you can cut, nail and torch@0xRhinoX | 17 min | $1.50 | — | Author’s pick | X post |
| 3D bathyscaphe with skills disabled@kenonews | — | $1.08 | — | — | X post |
| 3D racing game, same 64K tokens@gojoruns | 6.7 min | — | — | Author’s pick | X post |
| Rainy train station in 3D@haolv511482 | — | $3.91 | — | — | X post |
| Soft-body gummy eyeball in Three.js@daveeyyy_ | 8 min | $0.24 | — | — | X post |
| Personal site cover, same harness@jackyang | — | — | — | The poster picks Claude Opus 5.5; Kimi K3.1 Pro is good but plays it safe, and GPT-6 Astra is good but too abstract. | X post |
| Roadster animation, same steps@the0xbt | — | $2.00 | — | Author’s pick | X post |
| Mount Ember voxel volcano, Opus vs Kimi K3.1 Pro@semichenkko | — | $12.67 | — | — | X post |
| Blender scene recreation from references@notjazii | 25 min | $1.50 | — | — | X post |
| Haunted ship, one try each@olezhavich | 40 min | — | — | Kimi K3.1 Pro went for glowing waves, Claude Opus 5.5 for dark atmosphere and GPT-6.1 Sol for a cleaner ship; the poster picks Claude Opus 5.5 for atmosphere. | X post |
| Pelican on a bicycle, Kimi K3.1 Pro vs Kimi K3@hiwise_aigc | — | — | — | Author’s pick | X post |
| Motion design at highest reasoning, $86.50 vs $7.30@notjazii | 45 min | $7.30 | — | The poster says GPT-6 Astra is getting outdone by Kimi K3.1 Pro here, and leaves the final call to viewers. | X post |
| Microducks scene, high effort@BuildFastWithAI | 13 min | $1.57 | — | The poster says Claude Opus 5.5 still catches more details, but Kimi K3.1 Pro gets surprisingly close at about 3.5x less. | X post |
| 3D gamepad model with exploded view@leo114119 | — | — | — | — | X post |
| 12-step Rube Goldberg machine in 30 seconds@adolandev | — | — | — | — | X post |
| Orange orchard island with rain and day/night@_littlemandarin | 1 h 10 min | $4.00 | — | — | X post |
| Pixel airport scene, $0.98 vs $17.61@vib3coded | — | $0.98 | — | Author’s pick | X post |
| 3D pinball machine, first round@leppppp1 | 16.3 min | — | — | — | X post |
| Holographic CD spinning on its edge until it falls@Terry_Djony | — | — | — | — | X post |
| Ant colony foraging benchmark, four models@AlicanKiraz0 | — | $3.001 | — | — | X post |
| Interactive 3D stained-glass butterfly@ivanainai | — | — | — | Author’s pick | X post |
| Rocket launch simulator in one HTML file@asklogai | 23 min | $1.57 | — | — | X post |
| Sailing ship at sunset: Kimi K3.1 Pro vs Claude Opus 5.5@Da7_Tech | — | — | — | — | X post |
| Voxel SpongeBob in the Krusty Krab@MartinSzerment | 12 min | $0.54 | 54,000 | — | X post |
| 28-second dragon flight in six shots@kartikktwt | — | — | — | — | X post |
| Voxel kingdom world, Kimi K3.1 Pro vs Gemini 4 Carbon@lyraxana | — | — | — | — | X post |
| Pelican learns to ride a bike with RL@YifeiZhou | — | — | — | By the poster’s own eval, GPT-6 Astra scored 98/100 and Kimi K3.1 Pro 29/100. | X post |
| Full browser Minecraft, $2.30 vs $13.40@starmexxx | — | $2.30 | — | — | X post |
| Voxel ship in a bottle, $0.21 vs $9.81@vib3coded | — | $0.21 | — | Author’s pick | X post |
| “Lantern Hour” in Canvas 2D only@adolandev | — | — | — | — | X post |
| Animated promo SVG for a website@immadiuttej | 4 min | — | — | Author’s pick | X post |
| Kimi K3.1 Pro and Claude Opus 5.5, no details given@0x0SojalSec | — | — | — | — | X post |
| Animated 3D tabletop railway@Amank1412 | 18 min | $6.20 | — | — | X post |
| White House 3D website from the same images@MaxForAI | — | — | — | The poster says GPT-6 Astra is clearly better, while calling Kimi K3.1 Pro the open-source state of the art. | X post |
| Purple-palette motion piece, left vs right@OGKAIM | — | — | — | — | X post |
| Mechanical pelican riding a bike, Kimi K3.1 Pro solo@FiniYang | — | — | — | — | X post |
| Single-file Three.js test, three frontier models@AI_Screening | — | — | — | The poster says Kimi K3.1 Pro beat GPT-6 Astra and is now competing at the frontier. | X post |
| Animated voxel pagoda island@adxtyahq | 15 min | $5.71 | — | — | X post |
| Motion and detail test at top effort@kqyno7 | — | — | — | The poster says Kimi K3.1 Pro was beaten on every count and half-jokes that it was routed to a fake K3.1. | X post |
| Moon UI animation, Kimi K3.1 Pro vs Kimi K3@HarshithLucky3 | — | — | — | Author’s pick | X post |
| 3D pelican riding a bicycle, Kimi K3.1 Pro@chetaslua | — | — | — | — | X post |
| Recursive pelican scene, Kimi K3.1 Pro vs Claude Fable 5.5@chetaslua | — | — | — | — | X post |
Only numbers stated in each post. “—” means the post doesn’t say.
Who won where
44- Pixel city skyline, Kimi K3.1 Pro vs GPT-6.1 TerraThe poster says Terra was 5x faster at under 1/7 of the cost, but Kimi still looks better overall.— the author’s view, @AndyL5cc (paraphrased)
- Japanese pixel-art street corner, Kimi vs OpusThe poster says Kimi K3.1 Pro is a cheaper alternative, though not yet at Opus 5.5 quality.— the author’s view, @gmi_cloud (paraphrased)
- Voxel canal boat ride, Sonnet vs Kimi K3.1 ProThe poster thinks Claude Sonnet 5.5 won; Kimi K3.1 Pro got there after a few iterations and is not bad, but you cannot iterate your way into taste.— the author’s view, @filicroval (paraphrased)
- Meteor strike on a voxel island, MiniMax vs KimiThe poster says they paid 29x more for the worse one: MiniMax M3.1 Flash has bouncing debris and a treeline that burns tree by tree, Kimi K3.1 Pro has chunky cubes and a shapeless smoke column.— the author’s view, @semichenkko (paraphrased)
- Fart-powered backflip, three modelsThe poster says DeepSeek almost matched Claude Opus 5.5 for 24 times less, and that Kimi K3.1 Pro disappoints again.— the author’s view, @LatentOperatorX (paraphrased)
- Toy poodle model, code onlyThe poster expected more but finds both cute, and says Kimi K3.1 Pro tends to overthink and burn reasoning tokens first.— the author’s view, @harupm55 (paraphrased)
- Erupting 3D volcano, Kimi K3.1 Pro and 3 othersThe poster says Kimi K3.1 Pro has great detail and a big jump in 3D modeling, better than Astra Max.— the author’s view, @criscxuan (paraphrased)
- Moonwalking dancer under a spotlightThe poster names Kimi K3.1 Pro the winner: slow but efficient.— the author’s view, @LatentOperatorX (paraphrased)
- Jelly prism that splits real lightThe poster picks Kimi K3.1 Pro: one clean rainbow like the album cover, while Claude Opus 5.5 went full laser show with beams everywhere.— the author’s view, @0xRhinoX (paraphrased)
- Lighthouse in a storm, one HTML file, first tryThe poster calls the Kimi K3.1 Pro result garbage next to DeepSeek V4.1 Flash.— the author’s view, @Vortlyn (paraphrased)
- Short Three.js animation from scratch, four modelsThe poster ranks Claude Sonnet 5.5 first, then Gemini 4 Argon, Kimi K3.1 Pro and Muse Spark 1.3.— the author’s view, @VulKan42069 (paraphrased)
- Mechanical watch assembly in 3DThe poster says Kimi K3.1 Pro had all the time in the world and still lost to Claude Opus 5.5.— the author’s view, @kairox36 (paraphrased)
- Floatplane over a fjord, three modelsThe poster says Kimi K3.1 Pro on Max could not beat DeepSeek V4.1 Flash on xhigh, at 49 times the price.— the author’s view, @EnvolDev (paraphrased)
- Real-time Kerr black hole observatoryThe poster says Qwen 3.8 Flash Next rendered a real lensed black hole with overlays and readouts, while Kimi K3.1 Pro rendered something broken and unpolished.— the author’s view, @filicroval (paraphrased)
- Haunted carousel you can rideThe poster picks Kimi K3.1 Pro and asks whether Claude Opus 5.5 was nerfed.— the author’s view, @dromexa (paraphrased)
- Interactive 3D V8 engine lab in Three.jsThe poster says both work; Kimi K3.1 Pro has the more polished UI and a combustion panel, DeepSeek V4.1 Flash a more detailed dyno cell.— the author’s view, @sai0xgen (paraphrased)
- 3D solar system, four models, one shotThe poster says Kimi K3.1 Pro built the most polished UI of the four and also cost the most.— the author’s view, @nexightai (paraphrased)
- 15-second LLM showreel in codeThe poster calls DeepSeek V4.1 Flash 1017 the budget pick and says Kimi K3.1 Pro has bolder type, richer colors and more scene variety.— the author’s view, @EvoLinkAi (paraphrased)
- Six-legged robot modelThe poster says Kimi K3.1 Pro made a rounded shell with an inset lid, while GPT-6 Astra stacked a hexagonal slab on six boxy legs.— the author’s view, @0x_fokki (paraphrased)
- Sandworm model, Kimi vs AstraThe poster says Kimi K3.1 Pro covered every ring in panels and ran a gold band around the mouth, and prefers its result.— the author’s view, @elvar_official (paraphrased)
- 3D bee model, Kimi vs AstraThe poster says Kimi K3.1 Pro added hair on every leg and a faceted eye, while GPT-6 Astra built a smooth green plastic bee.— the author’s view, @Nyxera0x (paraphrased)
- Underwater research station and bathyscapheThe poster prefers Claude Opus 5.5's scene and asks whether he is doing something wrong with Kimi K3.1 Pro.— the author’s view, @kenonews (paraphrased)
- 3D lighthouse scene in the browserThe poster says Kimi K3.1 Pro went for rock detail, foam and clouds, and Claude Opus 5.5 for light, gulls and lightning, and asks which to ship.— the author’s view, @X1Kafadzhy (paraphrased)
- Black hole and pirate ship, Kimi vs OpusThe poster finds them hard to tell apart and thinks Kimi K3.1 Pro can challenge Claude.— the author’s view, @manmanloveai (paraphrased)
- US vs Iran chess, Gemini 4 Carbon vs Kimi K3.1 ProThe poster says Gemini 4 Carbon is clearly outperforming Kimi K3.1 Pro.— the author’s view, @abhinavflac (paraphrased)
- Foldable iPhone browser buildThe poster says it is not even close: Claude Opus 5.5 closes into a phone with a lock screen, while Kimi K3.1 Pro closes into a grey slab.— the author’s view, @cop_on_fire (paraphrased)
- Alpine Gold scene, max vs xhighThe poster says Kimi K3.1 Pro did better than on their Lighthouse test but is still far from Claude Opus 5.5.— the author’s view, @EnvolDev (paraphrased)
- Rainbow jelly you can cut, nail and torchThe poster praises Kimi K3.1 Pro’s colors and colored shadows and says Claude Opus 5.5 took 3x longer and cost 5x more.— the author’s view, @0xRhinoX (paraphrased)
- 3D racing game, same 64K tokensThe poster says Kimi K3.1 Pro shipped a working neon racer, while Claude Opus 5.5 never wrote a line.— the author’s view, @gojoruns (paraphrased)
- Personal site cover, same harnessThe poster picks Claude Opus 5.5; Kimi K3.1 Pro is good but plays it safe, and GPT-6 Astra is good but too abstract.— the author’s view, @jackyang (paraphrased)
- Roadster animation, same stepsThe poster says the $2 one looks better.— the author’s view, @the0xbt (paraphrased)
- Haunted ship, one try eachKimi K3.1 Pro went for glowing waves, Claude Opus 5.5 for dark atmosphere and GPT-6.1 Sol for a cleaner ship; the poster picks Claude Opus 5.5 for atmosphere.— the author’s view, @olezhavich (paraphrased)
- Pelican on a bicycle, Kimi K3.1 Pro vs Kimi K3The poster was surprised by how weak the Kimi K3 result was.— the author’s view, @hiwise_aigc (paraphrased)
- Motion design at highest reasoning, $86.50 vs $7.30The poster says GPT-6 Astra is getting outdone by Kimi K3.1 Pro here, and leaves the final call to viewers.— the author’s view, @notjazii (paraphrased)
- Microducks scene, high effortThe poster says Claude Opus 5.5 still catches more details, but Kimi K3.1 Pro gets surprisingly close at about 3.5x less.— the author’s view, @BuildFastWithAI (paraphrased)
- Pixel airport scene, $0.98 vs $17.61For value the poster picks Kimi K3.1 Pro, saying it holds its own next to Astra at almost 18 times less.— the author’s view, @vib3coded (paraphrased)
- Interactive 3D stained-glass butterflyThe poster says Kimi nailed the glass, lighting and gold details, while the Opus 5.5 result looked like a cheaper model.— the author’s view, @ivanainai (paraphrased)
- Pelican learns to ride a bike with RLBy the poster’s own eval, GPT-6 Astra scored 98/100 and Kimi K3.1 Pro 29/100.— the author’s view, @YifeiZhou (paraphrased)
- Voxel ship in a bottle, $0.21 vs $9.81Sol has more detail, but the poster picks Kimi K3.1 Pro and would not pay that much more for the extra detail.— the author’s view, @vib3coded (paraphrased)
- Animated promo SVG for a websiteThe poster picks Kimi K3.1 Pro for a cleaner design, better branding and smoother animation.— the author’s view, @immadiuttej (paraphrased)
- White House 3D website from the same imagesThe poster says GPT-6 Astra is clearly better, while calling Kimi K3.1 Pro the open-source state of the art.— the author’s view, @MaxForAI (paraphrased)
- Single-file Three.js test, three frontier modelsThe poster says Kimi K3.1 Pro beat GPT-6 Astra and is now competing at the frontier.— the author’s view, @AI_Screening (paraphrased)
- Motion and detail test at top effortThe poster says Kimi K3.1 Pro was beaten on every count and half-jokes that it was routed to a fake K3.1.— the author’s view, @kqyno7 (paraphrased)
- Moon UI animation, Kimi K3.1 Pro vs Kimi K3The poster finds the cute little moon animation very good.— the author’s view, @HarshithLucky3 (paraphrased)
Matchups with Kimi K3.1 Pro
17Claude Opus 5.5 vs Kimi K3.1 Pro52 posts · 316K viewsGPT-6 Astra vs Kimi K3.1 Pro25 posts · 241.1K viewsDeepSeek V4.1 Flash 1017 vs Kimi K3.1 Pro15 posts · 21.2K viewsGPT-6.1 Sol vs Kimi K3.1 Pro12 posts · 37.9K viewsGPT-6.1 Terra vs Kimi K3.1 Pro12 posts · 10.2K viewsKimi K3.1 Pro vs MiniMax M3.1 Flash5 posts · 13.4K viewsClaude Sonnet 5.5 vs Kimi K3.1 Pro5 posts · 7.9K viewsKimi K3 vs Kimi K3.1 Pro5 posts · 3.3K viewsGemini 4 Argon vs Kimi K3.1 Pro5 posts · 899 viewsGemini 4 Carbon vs Kimi K3.1 Pro2 posts · 19.8K viewsKimi K3.1 Pro vs Muse Spark 1.32 posts · 3.6K viewsClaude Fable 5.5 vs Kimi K3.1 Pro1 post · 186K viewsKimi K3.1 Pro vs Muse Spark Open1 post · 8K viewsKimi K3.1 Pro vs Qwen 3.8 Flash Next1 post · 1.5K viewsGLM 5.3 Flash vs Kimi K3.1 Pro1 post · 236 viewsGPT-6 Luna vs Kimi K3.1 Pro1 post · 85 viewsClaude Fable 5.1 vs Kimi K3.1 Pro1 post · 26 views
Spellings we map to Kimi K3.1 Pro
“kimi-k3.1-pro”, “Moonshot AI Kimi K3.1 Pro”, “K3.1 Pro”