正在加载视频...

视频加载失败

加载此视频时出现问题。这可能是由于临时网络问题，或视频可能不可用。

DeepSeek v4.1 Flash vs GLM 5.3 Flash I asked both models to recreate the Psychonauts 2 and Metroid Prime game menus as closely as possible to the originals. Attached are full-sized videos from both models for both games. All were done in a single shot, both at Max thinking,... show more

Mia

34,378 subscribers

23,835 次观看 • 10 天前 •via X (Twitter)

Anya Rossi• Live Now

Private livecam show

41 条评论

Ivan Fioravanti 的头像

Ivan Fioravanti10 天前

Yes GLM 5.3 Flash by far!

Mia 的头像

Mia10 天前

Yep, not even close !

Mia 的头像

Mia10 天前

The reference videos I gave both models:

Herlon Aguiar 的头像

Herlon Aguiar10 天前

Slowly we will see the hype of DS4.1F go away 🙂

filipe 的头像

filipe10 天前

i'm surprised, ds4.1 flash is actually good on this benchmark

Mia 的头像

Mia10 天前

done terribly on this one

Dan Harper 的头像

Dan Harper10 天前

GLM 5.3 Flash way better. Will have to try it

zoltan 的头像

zoltan10 天前

Deepseek is always so overly verbose. 1M ctx window on ds is equivalent to ~300k on glm. I love both, ds is way more fun to talk to but i like glm for execution.

Fatih Yeşilova 的头像

Fatih Yeşilova10 天前

Both models run on 4xspark? or for testing on cloud?

Mia 的头像

Mia10 天前

GLM on 2 sparks, DS on 3

Christopher Owen 的头像

Christopher Owen9 天前

Have you tried the harnesses from the model providers? maybe there is some secret sauce in zcode or deepcode/harness?

J A Z I I 的头像

J A Z I I10 天前

jeez why is v4.1 flash using more tokens? it's cheap after all right?

Mia 的头像

Mia10 天前

It thinks way too much and over complicates things, and in the end the results are worse.

J A Z I I 的头像

J A Z I I10 天前

yeah that's the worst thing about v4.1

Young Sneed 的头像

Young Sneed10 天前

its jus ta game menu

Vassio 的头像

Vassio10 天前

A quick question, what harnes did you use?

Mia 的头像

Mia10 天前

My own harness. Unreleased.

Atma Blabber 的头像

Atma Blabber10 天前

such test results these days are pretty much useless if you dont use the HARNESS that it is meant to be used with. glm with Zcode ds4.1 with DSH These results are with ur custom harness, and therefore are not only inaccurate but useless

Belcebuu 的头像

Belcebuu10 天前

Not very impressed by 4.1 at all such a disappointment

Mia 的头像

Mia10 天前

On these cases, very disappointed

Belcebuu 的头像

Belcebuu10 天前

It doubled the size and I don't see a big jump vs much smaller size ones like GLM 5.3 flash or Qwen3.8 next flash

Uncle cat bangkok 的头像

Uncle cat bangkok10 天前

Thats amazing!

hi 的头像

hi10 天前

Hory chet

Random Libertarian Tech Lead 的头像

Random Libertarian Tech Lead10 天前

I don't know, but I also know that this is a pretty worthless benchmark. Can we get serious about real useful benchmarks, rather than these worthless "do a bad one-shot of something useless, testing arbitrary trivial world knowledge" types of tests?

Inference User 的头像

Inference User10 天前

Putting both recreations side by side at max thinking is a clean comparison; identical prompts and retry counts would make the result even easier to reproduce.

PS5_BRO 的头像

PS5_BRO10 天前

GLM 5.3 Flash FTW

Egret 的头像

Egret10 天前

whats the harness

Paul 的头像

Paul10 天前

Have you tried the same prompt with qwen3.8 flash next? I gave it the task to make the koi-pond, using your prompt, and it did quite well.

Leo 的游戏人生的头像

Leo 的游戏人生10 天前

Token count is a cost, not a score. Burning 3–4x to land closer to the original isn't losing — it's paying more. Fidelity needs a rating, not a tally.Ran DS4.1F on real coding work this week. The token burn is real — but so is what comes out. Efficiency and capability aren't the same axis.

Kennan 的头像

Kennan10 天前

Sure DSF4.1 used far more tokens, but it got far better looking results in my opinion(havn't played the games).

MASA 的头像

MASA10 天前

Easily the lighter one did the other just overthink everything?

Paul Cleverly 的头像

Paul Cleverly10 天前

This type of example really brings into question just chasing the t/sec. Glm might be slower but like faster even at lower t/sec should probably include time taken with these comparisons.

Webster | JARVIS 的头像

Webster | JARVIS10 天前

GLM’s 3-4x token win matters. But menus are visual state machines, not text blobs. If 134k nails hover/transition fidelity, that’s a real architecture edge, not just brevity.

Carlo Pires 的头像

Carlo Pires10 天前

That's my perception for complex coding tasks too.

jeremy.og 的头像

jeremy.og10 天前

Better than qwen3.8-flash-next?

momo 🇺🇸 的头像

momo 🇺🇸10 天前

Both look really bad

basedcapital 的头像

basedcapital10 天前

a menu is all values, so every thinking step is another chance to swap one it saw for one it expected. whoever invented less won.

Dan C. 的头像

Dan C.10 天前

I'm starting to think DS results are better with supervision/well defined tasks and less so for open ended problems.

AtxSal 的头像

AtxSal10 天前

I think what I seen was leave my GLM build alone and wait for DS F 4.2.

0x7z7z 的头像

0x7z7z10 天前

I love these comparisons please make more of these!

Thomas Klein 的头像

Thomas Klein9 天前

I love seeing this stuff - but 90% of real world coding is fixing bugs / looking through repos, etc. not One shotting random websites and games. no real world engineer has stuff like this - they deal with edge cases, etc. thats what I am intersted in

相关视频

Qwen3.8-Flash-Next is shockingly good against GLM-3.5 Flash i used the same prompt to build fully controllable hummingbird using three.js and here's the side by side comparison between both it's insane how both are open models that can one shot entire game with a single prompt

Qwen3.8-Flash-Next is shockingly good against GLM-3.5 Flash i used the same prompt to build fully controllable hummingbird using three.js and here's the side by side comparison between both it's insane how both are open models that can one shot entire game with a single prompt

Tim Jayas

20,467 次观看 • 1 个月前

GLM 5.3 Flash vs Hybrid (GLM 5.3 Flash + Jev) on a turn-based strategy game (Halite 2) The hybrid approach is 13x faster, 56% of the pure GLM 5.3 Flash API cost, and gains a slight edge in performance. In the hybrid approach: -GLM 5.3 Flash handles strategic judgment of which planets matter, who expands, who fights, when to re-plan. -Jev handles the ship-by-ship execution. GLM 5.3 Flash alone defeats Jev alone 82% of the time. This is obviously just a single example and some of the first things I've tried to make use of Jev as an augmentation to an LLM. Could potentially make even more gains in cost, speed, and performance, but using what I've figured out here, I am compelled to try some ideas with robotics next.

GLM 5.3 Flash vs Hybrid (GLM 5.3 Flash + Jev) on a turn-based strategy game (Halite 2) The hybrid approach is 13x faster, 56% of the pure GLM 5.3 Flash API cost, and gains a slight edge in performance. In the hybrid approach: -GLM 5.3 Flash handles strategic judgment of which planets matter, who expands, who fights, when to re-plan. -Jev handles the ship-by-ship execution. GLM 5.3 Flash alone defeats Jev alone 82% of the time. This is obviously just a single example and some of the first things I've tried to make use of Jev as an augmentation to an LLM. Could potentially make even more gains in cost, speed, and performance, but using what I've figured out here, I am compelled to try some ideas with robotics next.

Harrison Kinsley

53,159 次观看 • 5 天前

how did DeepSeek-v4.1-Flash became so good both the outputs are from same prompts.. DeepSeek's flash model on max effort smokes Opus 5 on medium effort and the crazy part is it's still 90% cheaper best strategy would be to use Opus for planning and deepseek-v4.1-flash for execution, that would be an insane combo but do you think Gemini 4 Pro could outperform DeepSeek-v4.1-flash?

how did DeepSeek-v4.1-Flash became so good both the outputs are from same prompts.. DeepSeek's flash model on max effort smokes Opus 5 on medium effort and the crazy part is it's still 90% cheaper best strategy would be to use Opus for planning and deepseek-v4.1-flash for execution, that would be an insane combo but do you think Gemini 4 Pro could outperform DeepSeek-v4.1-flash?

Tim Jayas

11,371 次观看 • 9 天前

GLM 5.3 Flash vs Kimi K3 tested both models on same prompt at highest reasoning available but results came out really different > 5.3 flash took 20 minutes to complete this and costed $0.7 via AIhubMix API > kimi k3 took 30 minutes to complete this and costed $4.32 via Kimi Plan which one do you think did better here?

GLM 5.3 Flash vs Kimi K3 tested both models on same prompt at highest reasoning available but results came out really different > 5.3 flash took 20 minutes to complete this and costed $0.7 via AIhubMix API > kimi k3 took 30 minutes to complete this and costed $4.32 via Kimi Plan which one do you think did better here?

J A Z I I

23,105 次观看 • 1 个月前

deepseek v4.1 flash vs secret model tested both with the same prompt at the highest reasoning available > v4.1 flash took 2 hours to finish both tests and cost $2.60 > secret model finished both in 30 minutes, can’t reveal cost yet v4.1 flash spent a lot of time overthinking and running random tests, which slowed it down thanks AIHubmix for providing api access so i could test v4.1 flash hoping that team will cut down the overthinking before the full release, but for now here are the results which model did better here?

deepseek v4.1 flash vs secret model tested both with the same prompt at the highest reasoning available > v4.1 flash took 2 hours to finish both tests and cost $2.60 > secret model finished both in 30 minutes, can’t reveal cost yet v4.1 flash spent a lot of time overthinking and running random tests, which slowed it down thanks AIHubmix for providing api access so i could test v4.1 flash hoping that team will cut down the overthinking before the full release, but for now here are the results which model did better here?

J A Z I I

29,132 次观看 • 18 天前

glm 5.3 vs qwen 3.8 vs gemini 3.7 vs deepseek v4 flash four models designed and built three structures each on a physics-backed site, with no dimensions anywhere in the brief the setup: our own agent loop on OpenRouter, a construction site as the tool set – footings, walls, arches, roofs, scaffold, a lamp. the site enforces physics and nothing else: unsupported brick falls, a roof needs walls under it, a worker reaches 3.2 m above whatever he stands on, an arch needs centring until the keystone is set, concrete cures before it carries. no budget ceiling – material cost is tallied and reported, never blocked. tasks: 1. house – a plot and a palette, no plan. shape, height and material are the model's call 2. lighthouse – a headland cut by a gully, with a rock stack standing 30 m offshore. the lamp must burn, it must be the highest thing built, and the keeper must be able to walk to it 3. bridge – a river with one islet and banks at different heights. cross it however you want models: Z.ai glm 5.3 flash, Qwen qwen 3.8 flash, Google DeepMind gemini 3.7 flash, DeepSeek v4 flash vision all twelve objects were finished and signed off by the models themselves. tallest lighthouse is qwen's at 38.4 m, planted on the offshore stack with a bridge run out to it – the only model that read the site that way. deepseek signed off its bridge on an empty riverbed: 0 bricks, 107 minutes, $1.16m of material tallied - total cost, three builds #1 glm 5.3 flash – $0.201 #2 gemini 3.7 flash – $0.871 #3 qwen 3.8 flash – $1.058 #4 deepseek v4 flash – $1.567 - wall clock, three builds #1 gemini 3.7 flash – 91m #2 glm 5.3 flash – 228m #3 deepseek v4 flash – 502m #4 qwen 3.8 flash – 912m - total tokens #1 gemini 3.7 flash – 3,567,052 #2 glm 5.3 flash – 4,732,748 #3 qwen 3.8 flash – 13,469,333 #4 deepseek v4 flash – 18,230,076 - defects logged by the site #1 deepseek v4 flash – 59 #2 gemini 3.7 flash – 132 #3 glm 5.3 flash – 221 #4 qwen 3.8 flash – 350 - material tallied across three builds #1 gemini 3.7 flash – $359,884 #2 glm 5.3 flash – $583,358 #3 deepseek v4 flash – $1,327,484 #4 qwen 3.8 flash – $2,188,625 observations: • glm is the cheap one and nothing here is close – $0.201 for three buildings, $0.042 per million tokens, 6x under gemini's rate • what glm spends it on is bulk, not care: 166,228 bricks in one house and 156 defect weight, the worst single object in the set • gemini is the efficiency line – 91 minutes and 3.57m tokens for all three and an eighth of qwen's clock • gemini also builds the smallest of everything. its lighthouse is 22.5 m against qwen's 38.4, its house 6.9 m against 19.3 • qwen is the maximalist: 1.18m bricks, $2.19m of material, tallest on all three tasks, and 912 minutes – 15 hours – to get there conclusion: twelve finished objects for $3.80 all in, and a 7.8x price spread between the cheapest model and the priciest! follow thehype. for 24/7 ai news, analysis and breakdowns

glm 5.3 vs qwen 3.8 vs gemini 3.7 vs deepseek v4 flash four models designed and built three structures each on a physics-backed site, with no dimensions anywhere in the brief the setup: our own agent loop on OpenRouter, a construction site as the tool set – footings, walls, arches, roofs, scaffold, a lamp. the site enforces physics and nothing else: unsupported brick falls, a roof needs walls under it, a worker reaches 3.2 m above whatever he stands on, an arch needs centring until the keystone is set, concrete cures before it carries. no budget ceiling – material cost is tallied and reported, never blocked. tasks: 1. house – a plot and a palette, no plan. shape, height and material are the model's call 2. lighthouse – a headland cut by a gully, with a rock stack standing 30 m offshore. the lamp must burn, it must be the highest thing built, and the keeper must be able to walk to it 3. bridge – a river with one islet and banks at different heights. cross it however you want models: Z.ai glm 5.3 flash, Qwen qwen 3.8 flash, Google DeepMind gemini 3.7 flash, DeepSeek v4 flash vision all twelve objects were finished and signed off by the models themselves. tallest lighthouse is qwen's at 38.4 m, planted on the offshore stack with a bridge run out to it – the only model that read the site that way. deepseek signed off its bridge on an empty riverbed: 0 bricks, 107 minutes, $1.16m of material tallied - total cost, three builds #1 glm 5.3 flash – $0.201 #2 gemini 3.7 flash – $0.871 #3 qwen 3.8 flash – $1.058 #4 deepseek v4 flash – $1.567 - wall clock, three builds #1 gemini 3.7 flash – 91m #2 glm 5.3 flash – 228m #3 deepseek v4 flash – 502m #4 qwen 3.8 flash – 912m - total tokens #1 gemini 3.7 flash – 3,567,052 #2 glm 5.3 flash – 4,732,748 #3 qwen 3.8 flash – 13,469,333 #4 deepseek v4 flash – 18,230,076 - defects logged by the site #1 deepseek v4 flash – 59 #2 gemini 3.7 flash – 132 #3 glm 5.3 flash – 221 #4 qwen 3.8 flash – 350 - material tallied across three builds #1 gemini 3.7 flash – $359,884 #2 glm 5.3 flash – $583,358 #3 deepseek v4 flash – $1,327,484 #4 qwen 3.8 flash – $2,188,625 observations: • glm is the cheap one and nothing here is close – $0.201 for three buildings, $0.042 per million tokens, 6x under gemini's rate • what glm spends it on is bulk, not care: 166,228 bricks in one house and 156 defect weight, the worst single object in the set • gemini is the efficiency line – 91 minutes and 3.57m tokens for all three and an eighth of qwen's clock • gemini also builds the smallest of everything. its lighthouse is 22.5 m against qwen's 38.4, its house 6.9 m against 19.3 • qwen is the maximalist: 1.18m bricks, $2.19m of material, tallest on all three tasks, and 912 minutes – 15 hours – to get there conclusion: twelve finished objects for $3.80 all in, and a 7.8x price spread between the cheapest model and the priciest! follow thehype. for 24/7 ai news, analysis and breakdowns

thehype.

26,360 次观看 • 29 天前

GLM-5.3 Flash in already nerfed? I compared Ox Alpha in it's prime vs today just before it got revealed as GLM-5.3 Flash using the same prompt to build 3d raptor engine using three.js OxAlpha on 24th August gave far more accurate output compared to today which looks way worse prompt i used: Create a 3d SpaceX raptor engine using three.js, use white theme background so the main object is clearly visible

GLM-5.3 Flash in already nerfed? I compared Ox Alpha in it's prime vs today just before it got revealed as GLM-5.3 Flash using the same prompt to build 3d raptor engine using three.js OxAlpha on 24th August gave far more accurate output compared to today which looks way worse prompt i used: Create a 3d SpaceX raptor engine using three.js, use white theme background so the main object is clearly visible

Tim Jayas

78,962 次观看 • 1 个月前

deepseek v4.1 flash ran all four three.js scenes for $0.15. gemini 3.8 flash needed $1.25 we put the two models on one job: a self-contained html page that renders an animated 3d scene with three.js – a storm at sea, a spiral galaxy, a tornado over farmland, a tesla coil on a dark table. same prompt - calls until all four scenes were usable #1 deepseek v4.1 flash – 6 #2 gemini 3.8 flash – 12 - time #1 deepseek v4.1 flash – 13m 57s #2 gemini 3.8 flash – 32m 02s - cost #1 deepseek v4.1 flash – $0.15 #2 gemini 3.8 flash – $1.25 observations: • gemini blew out the exposure on 4 of 5 first attempts – a white blob where the tesla coil should be. the prompt had a paragraph on exposure. it did not help. • both models overcorrect in the fix loop: "too bright" becomes "almost black". round two is where it lands. same four scenes, 6 calls and $0.15 against 12 calls and $1.25!

deepseek v4.1 flash ran all four three.js scenes for $0.15. gemini 3.8 flash needed $1.25 we put the two models on one job: a self-contained html page that renders an animated 3d scene with three.js – a storm at sea, a spiral galaxy, a tornado over farmland, a tesla coil on a dark table. same prompt - calls until all four scenes were usable #1 deepseek v4.1 flash – 6 #2 gemini 3.8 flash – 12 - time #1 deepseek v4.1 flash – 13m 57s #2 gemini 3.8 flash – 32m 02s - cost #1 deepseek v4.1 flash – $0.15 #2 gemini 3.8 flash – $1.25 observations: • gemini blew out the exposure on 4 of 5 first attempts – a white blob where the tesla coil should be. the prompt had a paragraph on exposure. it did not help. • both models overcorrect in the fix loop: "too bright" becomes "almost black". round two is where it lands. same four scenes, 6 calls and $0.15 against 12 calls and $1.25!

thehype.

62,603 次观看 • 16 天前

glm 5.3 flash is 7.5x cheaper, but 3.4x slower than gemini 3.7 flash Z.ai glm 5.3 flash – shipped aug 26, $0.07/$0.25 per 1m Google DeepMind gemini 3.7 flash – shipped aug 13, $0.38/$1.88 per 1m we put the two models on one job: write one html file that draws an animated 3d scene in the browser. no images, no downloads, and it has to look the same on every load. the setup: three scenes – a glass aquarium in a lit room, the solar system, a night city under a thunderstorm. identical brief word for word, reasoning effort high, 64k output cap. the numbers below are not the whole run. they cover the three scenes we kept – the best one per task from each model, the ones in the video. - total generation time for the three scenes #1 gemini 3.7 flash – 10m 36s #2 glm 5.3 flash – 36m 30s - tokens spent on those three scenes #1 glm 5.3 flash – 110k #2 gemini 3.7 flash – 111k - cost of those three scenes #1 glm 5.3 flash – $0.027 #2 gemini 3.7 flash – $0.202 observations: • glm's first 10 attempts: 7 blank pages. it kept inventing short random helpers and forgetting to define one of them. the fix was one line in the brief: use exactly one random helper, named rand(), and don't invent shorthands next to it. next 12 attempts: 11 alive, 0 crashes. • glm spends 66% of its output on reasoning, gemini 57%. that is the whole speed gap. • gemini's storm came back as a black rectangle in 4 of 6 runs. glm's best storm has a branching bolt, lit rain and wet asphalt – for $0.01. conclusion: same three scenes, same token spend – glm 5.3 flash billed $0.027 and took 36m 30s, gemini 3.7 flash billed $0.202 and took 10m 36s. glm wins gemini on price and made the best storm of the whole run follow thehype. for 24/7 ai news, analysis and breakdowns

glm 5.3 flash is 7.5x cheaper, but 3.4x slower than gemini 3.7 flash Z.ai glm 5.3 flash – shipped aug 26, $0.07/$0.25 per 1m Google DeepMind gemini 3.7 flash – shipped aug 13, $0.38/$1.88 per 1m we put the two models on one job: write one html file that draws an animated 3d scene in the browser. no images, no downloads, and it has to look the same on every load. the setup: three scenes – a glass aquarium in a lit room, the solar system, a night city under a thunderstorm. identical brief word for word, reasoning effort high, 64k output cap. the numbers below are not the whole run. they cover the three scenes we kept – the best one per task from each model, the ones in the video. - total generation time for the three scenes #1 gemini 3.7 flash – 10m 36s #2 glm 5.3 flash – 36m 30s - tokens spent on those three scenes #1 glm 5.3 flash – 110k #2 gemini 3.7 flash – 111k - cost of those three scenes #1 glm 5.3 flash – $0.027 #2 gemini 3.7 flash – $0.202 observations: • glm's first 10 attempts: 7 blank pages. it kept inventing short random helpers and forgetting to define one of them. the fix was one line in the brief: use exactly one random helper, named rand(), and don't invent shorthands next to it. next 12 attempts: 11 alive, 0 crashes. • glm spends 66% of its output on reasoning, gemini 57%. that is the whole speed gap. • gemini's storm came back as a black rectangle in 4 of 6 runs. glm's best storm has a branching bolt, lit rain and wet asphalt – for $0.01. conclusion: same three scenes, same token spend – glm 5.3 flash billed $0.027 and took 36m 30s, gemini 3.7 flash billed $0.202 and took 10m 36s. glm wins gemini on price and made the best storm of the whole run follow thehype. for 24/7 ai news, analysis and breakdowns

thehype.

15,997 次观看 • 1 个月前

Opus 4.8 vs GLM 5.2 everyone is praising Glm on my timeline, so i had to test it myself > both models were given the same prompt > GLM was super fast and used like 2% of my weekly limit > Opus took over 20 mins and used like 3% of weekly limits which one do you think won?

Opus 4.8 vs GLM 5.2 everyone is praising Glm on my timeline, so i had to test it myself > both models were given the same prompt > GLM was super fast and used like 2% of my weekly limit > Opus took over 20 mins and used like 3% of weekly limits which one do you think won?

J A Z I I

48,130 次观看 • 3 个月前

DeepSeek V4.1 Flash vs Gemini 3.8 Flash at 3D geometry test: six self-contained three.js stereometry scenes — a (3,8) torus knot, a Klein bottle, a great stellated dodecahedron, a hyperboloid ruled by 120 straight rods, a level-3 Menger sponge, a thick Möbius strip. one shot each, no libraries beyond three.js, every surface painted with a live iridescent rainbow that has to crawl across the geometry. cost: Gemini 3.8 Flash — $0.11 DeepSeek V4.1 Flash — $0.27 • gemini delivers all six figures for 2.4× less — 21k output tokens against deepseek's 168k • gemini is 5× faster — every scene under 25 seconds, deepseek took up to 3 minutes on the Möbius strip • deepseek wins the torus knot: its rope is vivid and glossy, while gemini rendered the same knot nearly black • neither model can build a real Klein bottle. both hand back a twisted tube and call it done try both on aimlapi[.]com

DeepSeek V4.1 Flash vs Gemini 3.8 Flash at 3D geometry test: six self-contained three.js stereometry scenes — a (3,8) torus knot, a Klein bottle, a great stellated dodecahedron, a hyperboloid ruled by 120 straight rods, a level-3 Menger sponge, a thick Möbius strip. one shot each, no libraries beyond three.js, every surface painted with a live iridescent rainbow that has to crawl across the geometry. cost: Gemini 3.8 Flash — $0.11 DeepSeek V4.1 Flash — $0.27 • gemini delivers all six figures for 2.4× less — 21k output tokens against deepseek's 168k • gemini is 5× faster — every scene under 25 seconds, deepseek took up to 3 minutes on the Möbius strip • deepseek wins the torus knot: its rope is vivid and glossy, while gemini rendered the same knot nearly black • neither model can build a real Klein bottle. both hand back a twisted tube and call it done try both on aimlapi[.]com

AI/ML API

39,857 次观看 • 16 天前

i burned through my entire GPT-6 Astra weekly quota in 3 days so i'm replacing my $200/mo Codex subscription with this... a free 32,960-star GitHub repo that runs GLM-5.3 Flash, Kimi K3, and DeepSeek V4 Flash locally [here's the exact setup i used:] 1. get colibri download the release archive for Linux, macOS, or Windows. no compiler needed. python3 coli info 2. or build it from source git clone cd colibri/c ./setup.sh cd .. 3. pick a model GLM-5.3 Flash: 321B, ~195 GB converted, 25 GB RAM DeepSeek V4 Flash: 284B, ~167 GB, or ~85 GB for the REAP 150B version Kimi K3: 2.8T, ~1.6 TB snapshot, 32 GB+ RAM none of them needs a GPU. a GPU only makes some runs faster. 4. get the weights GLM-5.3 Flash uses the dedicated converter: python3 tools/convert_glm53.py --outdir /nvme/glm53_i4 --min-free-gb 30 DeepSeek V4 Flash and Kimi K3 load from their original Hugging Face snapshots. pre-converted containers work too. 5. run it COLI_MODEL=/nvme/glm53_i4 ./coli chat ./coli web --model /nvme/glm53_i4 # dashboard ./coli serve --model /nvme/glm53_i4 # headless API 6. tune it ./coli doctor # readiness check ./coli plan # RAM/VRAM/disk placement ./coli tune # fastest safe profile `coli` reads `config.json` and picks the right engine for each model. the tradeoff is honest: no token bill, no hosted quota... but you need the storage. the frontier moved from your subscription to your SSD.

Avid

427,527 次观看 • 11 天前

Seedance 2.0 vs Omni Flash 🥊 Ngl, this side-by-side comparison is kinda wild. I took the exact same storyboard and ran it through both models to see how they handle movement and consistency. 🎨 Storyboard: GPT Image 2 🎬 Animation: Seedance 2.0 (Top) vs Omni Flash (Bottom) Tbh both models popped off, but the vibes are totally different. Are you team 🐶 (Seedance) or team 🐱 (Omni)? Let’s settle this in the replies! 👇 (Dropping the exact prompts in the thread 🧵)

Seedance 2.0 vs Omni Flash 🥊 Ngl, this side-by-side comparison is kinda wild. I took the exact same storyboard and ran it through both models to see how they handle movement and consistency. 🎨 Storyboard: GPT Image 2 🎬 Animation: Seedance 2.0 (Top) vs Omni Flash (Bottom) Tbh both models popped off, but the vibes are totally different. Are you team 🐶 (Seedance) or team 🐱 (Omni)? Let’s settle this in the replies! 👇 (Dropping the exact prompts in the thread 🧵)

Ivanna | AI Art & Prompts

23,642 次观看 • 3 个月前

Metroid Prime, running native and stable on iPhone and iPad for the first time. No emulator. No streaming from a Mac. This is EctoPad. Built on Metaforce, the native Metroid Prime engine developed by AxioDL and its contributors, it plays great on both devices and runs at full speed on iPhone 14 and newer. Touch controls, controller support, and 4x internal res. I am still working through a visual issue that causes some textures and lighting to render inaccurately, along with a smaller problem affecting the first tutorial door. Sol 5.6 was used to make this possible. Open sourcing it below.

Metroid Prime, running native and stable on iPhone and iPad for the first time. No emulator. No streaming from a Mac. This is EctoPad. Built on Metaforce, the native Metroid Prime engine developed by AxioDL and its contributors, it plays great on both devices and runs at full speed on iPhone 14 and newer. Touch controls, controller support, and 4x internal res. I am still working through a visual issue that causes some textures and lighting to render inaccurately, along with a smaller problem affecting the first tutorial door. Sol 5.6 was used to make this possible. Open sourcing it below.

Kahris

450,903 次观看 • 1 个月前

ox alpha vs deepseek v4 flash vision vs grok 4.6 vs gemini 3.7 flash vs – on photo-to-3d four vision models got one photograph each and had to rebuild the place inside it as a Three.js scene. twelve scenes, twelve first-try runs, zero console errors the setup: one reference photo per scene, sent as an image on OpenRouter. the prompt never says what is in the picture – no "motel", no "bar", no "gas station". the model has to read the photo and rebuild it: layout, materials, hour of the day, and whatever is around the corner that the frame does not show tasks – three photographs of early-2000s america: 1. a motel at night, neon pylon lit, snow on the ground 2. an old new york tavern interior, tin ceiling, tiled floor 3. an abandoned service station in the california desert, midday sun each scene ships as one self-contained html file, procedural geometry and canvas textures only, no downloads. three timed camera shots, and shot 1 has to reproduce the framing of the reference photo models: xAI grok 4.6, Google DeepMind gemini 3.7 flash, DeepSeek deepseek v4 flash vision exp, and ox alpha – a stealth model on openrouter, free, no lab attached to it yet results: - wall clock, three scenes #1 gemini 3.7 flash – 11m 12s #2 deepseek v4 flash – 15m 20s #3 grok 4.6 – 28m 11s #4 ox alpha – 38m 54s - output tokens #1 gemini 3.7 flash – 77,396 #2 ox alpha – 87,613 #3 grok 4.6 – 105,687 #4 deepseek v4 flash – 127,884 - lines of code shipped #1 ox alpha – 2,090 #2 deepseek v4 flash – 2,291 #3 grok 4.6 – 3,529 #4 gemini 3.7 flash – 3,989 - total price #1 ox alpha – $0.000 #2 deepseek v4 flash – $0.091 #3 gemini 3.7 flash – $0.136 #4 grok 4.6 – $0.697 observations: • grok is 7.7x the price of deepseek. it is the only model that read the light – low sun, real shadows on the station, a cold night on the motel • gemini is the fastest and the least deliberate. 17,158 reasoning tokens against deepseek's 99,172, and it still shipped the most code – 3,989 lines • deepseek thought hardest and rendered plainest. 99,172 reasoning tokens, 5.8x gemini's, spent on layout rather than on light. its motel is the second best in the set for $0.030 • ox alpha is free and reads a photo as well as anything here – it lifted "family units / kitchenettes" off the pylon and redrew it in canvas conclusion: twelve scenes, four models, zero fixes, and the whole run cost $0.924! follow thehype. for 24/7 ai news, analysis and breakdowns

ox alpha vs deepseek v4 flash vision vs grok 4.6 vs gemini 3.7 flash vs – on photo-to-3d four vision models got one photograph each and had to rebuild the place inside it as a Three.js scene. twelve scenes, twelve first-try runs, zero console errors the setup: one reference photo per scene, sent as an image on OpenRouter. the prompt never says what is in the picture – no "motel", no "bar", no "gas station". the model has to read the photo and rebuild it: layout, materials, hour of the day, and whatever is around the corner that the frame does not show tasks – three photographs of early-2000s america: 1. a motel at night, neon pylon lit, snow on the ground 2. an old new york tavern interior, tin ceiling, tiled floor 3. an abandoned service station in the california desert, midday sun each scene ships as one self-contained html file, procedural geometry and canvas textures only, no downloads. three timed camera shots, and shot 1 has to reproduce the framing of the reference photo models: xAI grok 4.6, Google DeepMind gemini 3.7 flash, DeepSeek deepseek v4 flash vision exp, and ox alpha – a stealth model on openrouter, free, no lab attached to it yet results: - wall clock, three scenes #1 gemini 3.7 flash – 11m 12s #2 deepseek v4 flash – 15m 20s #3 grok 4.6 – 28m 11s #4 ox alpha – 38m 54s - output tokens #1 gemini 3.7 flash – 77,396 #2 ox alpha – 87,613 #3 grok 4.6 – 105,687 #4 deepseek v4 flash – 127,884 - lines of code shipped #1 ox alpha – 2,090 #2 deepseek v4 flash – 2,291 #3 grok 4.6 – 3,529 #4 gemini 3.7 flash – 3,989 - total price #1 ox alpha – $0.000 #2 deepseek v4 flash – $0.091 #3 gemini 3.7 flash – $0.136 #4 grok 4.6 – $0.697 observations: • grok is 7.7x the price of deepseek. it is the only model that read the light – low sun, real shadows on the station, a cold night on the motel • gemini is the fastest and the least deliberate. 17,158 reasoning tokens against deepseek's 99,172, and it still shipped the most code – 3,989 lines • deepseek thought hardest and rendered plainest. 99,172 reasoning tokens, 5.8x gemini's, spent on layout rather than on light. its motel is the second best in the set for $0.030 • ox alpha is free and reads a photo as well as anything here – it lifted "family units / kitchenettes" off the pylon and redrew it in canvas conclusion: twelve scenes, four models, zero fixes, and the whole run cost $0.924! follow thehype. for 24/7 ai news, analysis and breakdowns

thehype.

24,272 次观看 • 1 个月前

GPT-5.6 LUNA AND DEEPSEEK V4 FLASH ARE BOTH FREE RIGHT NOW -- ONE IN A CHATBOT, THE OTHER IN A BROWSER IDE • ChatGPT, the chatbot > GPT-5.6 Luna is now the default on Free and Go: > Unlimited text chats and a Think button land next week. • MonkeyCode, the browser IDE > 30M tokens a day, resets daily, no card: > DeepSeek V4 Flash, Qwen, GLM, Kimi and MiniMax behind it. > Cloud workspace at 1 vCPU and 4GB, nothing to install. Luna posts 62% fewer factual errors than the model it replaces. MonkeyCode is open source from Chaitin and runs on their cloud, so aim it at side projects. One is live today -> The other arrives next week.

GPT-5.6 LUNA AND DEEPSEEK V4 FLASH ARE BOTH FREE RIGHT NOW -- ONE IN A CHATBOT, THE OTHER IN A BROWSER IDE • ChatGPT, the chatbot > GPT-5.6 Luna is now the default on Free and Go: > Unlimited text chats and a Think button land next week. • MonkeyCode, the browser IDE > 30M tokens a day, resets daily, no card: > DeepSeek V4 Flash, Qwen, GLM, Kimi and MiniMax behind it. > Cloud workspace at 1 vCPU and 4GB, nothing to install. Luna posts 62% fewer factual errors than the model it replaces. MonkeyCode is open source from Chaitin and runs on their cloud, so aim it at side projects. One is live today -> The other arrives next week.

slash1s

33,804 次观看 • 1 个月前

UPDATE: Charlie Kirk 🚨 Muzzle Flash: Second Shooter Location, Reflection, or Both? This video captures a flash on a window right as Charlie Kirk was shot. Let’s break this down… follow the numbers on the videos. #1. This video shows what appears to be a muzzle flash, just as Charlie is shot, leading people to believe the shooter was to Charlie’s right.. Notate the people to the left of the flash #2. This photo shows the location of the flash. Notate people to the left, on a middle-landing on a staircase. #3. This photo shows a clearer image showing both the middle-landing and the same location of the flash. It is a window. #4 . This video shows the building behind Charlie, which is a long hallway. This debunks any claims the shot came from here. — I zoom in where Charlie was — I zoom in on staircase — I zoom in on the window/flash — I zoom in on where suspected shooter was #5. I notate the flash reflection angles. — Video angle #1 notated — two reflection paths notated in relation to Video angle #1. 1. Reflection angle shows where we were told the shooter was. 2. Reflection angle shows mirrored angle, obviously where there is no reports of a shooter being in. 🔻 Final conclusion: This to me is very likely a muzzle flash reflection from a far off location. I am unsure of the reflection angle however, but this can be 100% proven if someone were to recreate the angles… maybe with a flash camera. If anyone is willing to, or has the means to do this (safely), this will either prove the shot came where the FBI said it came from, or it will prove there was a second shooter, in the direction of the mirrored angle.

UPDATE: Charlie Kirk 🚨 Muzzle Flash: Second Shooter Location, Reflection, or Both? This video captures a flash on a window right as Charlie Kirk was shot. Let’s break this down… follow the numbers on the videos. #1. This video shows what appears to be a muzzle flash, just as Charlie is shot, leading people to believe the shooter was to Charlie’s right.. Notate the people to the left of the flash #2. This photo shows the location of the flash. Notate people to the left, on a middle-landing on a staircase. #3. This photo shows a clearer image showing both the middle-landing and the same location of the flash. It is a window. #4 . This video shows the building behind Charlie, which is a long hallway. This debunks any claims the shot came from here. — I zoom in where Charlie was — I zoom in on staircase — I zoom in on the window/flash — I zoom in on where suspected shooter was #5. I notate the flash reflection angles. — Video angle #1 notated — two reflection paths notated in relation to Video angle #1. 1. Reflection angle shows where we were told the shooter was. 2. Reflection angle shows mirrored angle, obviously where there is no reports of a shooter being in. 🔻 Final conclusion: This to me is very likely a muzzle flash reflection from a far off location. I am unsure of the reflection angle however, but this can be 100% proven if someone were to recreate the angles… maybe with a flash camera. If anyone is willing to, or has the means to do this (safely), this will either prove the shot came where the FBI said it came from, or it will prove there was a second shooter, in the direction of the mirrored angle.

MJTruthUltra

10,665,164 次观看 • 1 年前

FREE DeepSeek V4 Pro and V4 Flash API. Runtime by Bad Theory Labs just dropped it. Same platform, new free model, same 2-minute setup. > Runtime already gave you free Claude Opus 4.8 and GPT 5.5 access. Now they added DeepSeek V4 Pro and V4 Flash to the free tier. Absolutely FREE, YES. Same exact flow as before: → Go to runtime(.)badtheorylabs(.)com → Sign up with Google → Fill in your details during onboarding → Free balance lands automatically(But you don't need it to use DeepSeek - it's completely free right now) → Grab your API key from the dashboard → Plug into OpenCode, OpenClaw, or Hermes - done DeepSeek V4 Pro is one of the strongest coding and reasoning models available right now. Free tier also includes btl-2 - auto-routes between GLM, OpenAI, Anthropic, and Gemma. 10 million tokens per month at zero cost. 344 models total. One OpenAI-compatible endpoint. No code rebuild, no card, no limits hunting. >> Just start using the API, and when choosing a provider, select DeepSeek I tested it myself - fully working. Bookmark this before the free tier fills up.

FREE DeepSeek V4 Pro and V4 Flash API. Runtime by Bad Theory Labs just dropped it. Same platform, new free model, same 2-minute setup. > Runtime already gave you free Claude Opus 4.8 and GPT 5.5 access. Now they added DeepSeek V4 Pro and V4 Flash to the free tier. Absolutely FREE, YES. Same exact flow as before: → Go to runtime(.)badtheorylabs(.)com → Sign up with Google → Fill in your details during onboarding → Free balance lands automatically(But you don't need it to use DeepSeek - it's completely free right now) → Grab your API key from the dashboard → Plug into OpenCode, OpenClaw, or Hermes - done DeepSeek V4 Pro is one of the strongest coding and reasoning models available right now. Free tier also includes btl-2 - auto-routes between GLM, OpenAI, Anthropic, and Gemma. 10 million tokens per month at zero cost. 344 models total. One OpenAI-compatible endpoint. No code rebuild, no card, no limits hunting. >> Just start using the API, and when choosing a provider, select DeepSeek I tested it myself - fully working. Bookmark this before the free tier fills up.

Atenov int.

27,487 次观看 • 3 个月前

Im done with GLM-5.3 Flash 🤯 I went deep into the guts of GLM-5.3 Flash. Ripped out every expert, every memory head, every quantization scale, looked at all of it up close, then put the model back together so you can see the inside too. 🎉 GLM-5.3 Flash NVFP4 is now in Weight Atlas. 320B total, 18B active, 45 language layers and a 24 block Vision tower. And none of it is read from the config. I ran the deployed NVFP4 checkpoint itself on 4x RTX PRO 6000 and measured what it does. - 12,096 expert cells, 42 layers x 288 experts, each with exact REAP importance from 12.59M tokens per layer, route share and output contribution. Slice it by 14 domains, russian and vision included - one chart invalidated an assumption I had. Route count does not track expert importance, Spearman -0.222 against exact REAP. The ranking itself is stable, 0.994 split half - 2,176 KDA memory heads, each with its own half life, from 0.25 tokens to 2.7M - 19 billion NVFP4 block scales, plus the deployed FC2 QDQ error: median 21.05 dB and 21.7% of values turn into exact zeros - a causal check of REAP at inference, weights untouched. Dropping the top 2% experts changes 79% of sequences, dropping the bottom 2% changes 64% Every number comes from a capture I checksummed, 15 artifacts, and the limits are written next to each chart. Sound on I will record a video walk through every chart later. For now go poke around inside it:

Im done with GLM-5.3 Flash 🤯 I went deep into the guts of GLM-5.3 Flash. Ripped out every expert, every memory head, every quantization scale, looked at all of it up close, then put the model back together so you can see the inside too. 🎉 GLM-5.3 Flash NVFP4 is now in Weight Atlas. 320B total, 18B active, 45 language layers and a 24 block Vision tower. And none of it is read from the config. I ran the deployed NVFP4 checkpoint itself on 4x RTX PRO 6000 and measured what it does. - 12,096 expert cells, 42 layers x 288 experts, each with exact REAP importance from 12.59M tokens per layer, route share and output contribution. Slice it by 14 domains, russian and vision included - one chart invalidated an assumption I had. Route count does not track expert importance, Spearman -0.222 against exact REAP. The ranking itself is stable, 0.994 split half - 2,176 KDA memory heads, each with its own half life, from 0.25 tokens to 2.7M - 19 billion NVFP4 block scales, plus the deployed FC2 QDQ error: median 21.05 dB and 21.7% of values turn into exact zeros - a causal check of REAP at inference, weights untouched. Dropping the top 2% experts changes 79% of sequences, dropping the bottom 2% changes 64% Every number comes from a capture I checksummed, 15 artifacts, and the limits are written next to each chart. Sound on I will record a video walk through every chart later. For now go poke around inside it:

Alexey Fateev

56,181 次观看 • 24 天前

NVIDIA is giving away free access to 130+ AI models for a full year > most people building AI agents are paying $50-200/month for API access NVIDIA just made that argument irrelevant models you get: MiniMax M2.7, GLM 5.1, Kimi 2.5, DeepSeek-v4-flash, GPT-OSS-120B and 110+ more setup: > step 1 - get your free key > go to > register -> bind phone -> copy API key > step 2 - add to Hermes agent > open Settings -> Model Provider -> Custom base_url = " api_key = "nvapi-xxxxxxxxxxxxxxxxxxxx" > step 3 - pick a model model = "minimaxai/minimax-m2.7" model = "zhipuai/glm-5.1" model = "moonshot-ai/kimi-2.5" model = "deepseek-ai/deepseek-v4-flash" model = "nvidia/nemotron-3-ultra-550b-a55b" > Hermes already has NVIDIA set as default base_url > paste the key and you're running instantly > works the same in Cursor and OpenCode > cost: $0 > limit: 40 req/min > expires: 1 year while everyone is paying for API access, this is sitting there for free

NVIDIA is giving away free access to 130+ AI models for a full year > most people building AI agents are paying $50-200/month for API access NVIDIA just made that argument irrelevant models you get: MiniMax M2.7, GLM 5.1, Kimi 2.5, DeepSeek-v4-flash, GPT-OSS-120B and 110+ more setup: > step 1 - get your free key > go to > register -> bind phone -> copy API key > step 2 - add to Hermes agent > open Settings -> Model Provider -> Custom base_url = " api_key = "nvapi-xxxxxxxxxxxxxxxxxxxx" > step 3 - pick a model model = "minimaxai/minimax-m2.7" model = "zhipuai/glm-5.1" model = "moonshot-ai/kimi-2.5" model = "deepseek-ai/deepseek-v4-flash" model = "nvidia/nemotron-3-ultra-550b-a55b" > Hermes already has NVIDIA set as default base_url > paste the key and you're running instantly > works the same in Cursor and OpenCode > cost: $0 > limit: 40 req/min > expires: 1 year while everyone is paying for API access, this is sitting there for free

Mr. Buzzoni

196,730 次观看 • 3 个月前