Loading video...

Video Failed to Load

There was a problem loading this video. This could be due to a temporary network issue or the video might be unavailable.

GLM-5.2 delivers a substantial leap in app development capabilities, which also represent demanding long-horizon tasks. Results: - GLM-5.1: 21/70 - GLM-5.2: 48/70 - Claude Fable 5: 56/70 That's more than a twofold improvement from GLM-5.1 to GLM-5.2. These come from an internal benchmark of 35 challenging mobile development tasks,... show more

Zixuan Li

20,685 subscribers

322,893 views • 1 month ago •via X (Twitter)

News & Politics Education Science & Technology

Anya Rossi• Live Now

Private livecam show

0 Comments

No comments available

Comments from the original post will appear here

Related Videos

Introducing GLM 5.2 for autoresearch GLM 5.2 is the first open weights model we've tried on our autoresearch pipeline that's proven capable for real research tasks. With Fable 5's restrictions on research, having an open weights alternative is a huge win for open source Watch it carry out fully async vs colocated sync RL training on Harbor code contests across two 8xH100 nodes on top of SkyRL. Resolves setup issues, tracks runs to completion, and produces a full comparison of throughput and reward stability

Introducing GLM 5.2 for autoresearch GLM 5.2 is the first open weights model we've tried on our autoresearch pipeline that's proven capable for real research tasks. With Fable 5's restrictions on research, having an open weights alternative is a huge win for open source Watch it carry out fully async vs colocated sync RL training on Harbor code contests across two 8xH100 nodes on top of SkyRL. Resolves setup issues, tracks runs to completion, and produces a full comparison of throughput and reward stability

alphaXiv

270,565 views • 1 month ago

Z.ai GLM 5.2 is now live on Eigent from day 0! threw a real long-horizon task: research 30 companies across 6 sectors of the AI infrastructure stack, structure it into JSON, then build an interactive HTML report. same prompt, 5.1 vs 5.2. where 5.2 pulls ahead: → plans deeper and verifies harder. → researches wider. Covers 2x the ground. → ships a more complete, interactive report. with a 1M context window, it holds the whole task from research to final deliverable. no dropped threads, no half-finished output. long tasks are where the difference shows. try it on open source coworker Eigent. 🦾

Z.ai GLM 5.2 is now live on Eigent from day 0! threw a real long-horizon task: research 30 companies across 6 sectors of the AI infrastructure stack, structure it into JSON, then build an interactive HTML report. same prompt, 5.1 vs 5.2. where 5.2 pulls ahead: → plans deeper and verifies harder. → researches wider. Covers 2x the ground. → ships a more complete, interactive report. with a 1M context window, it holds the whole task from research to final deliverable. no dropped threads, no half-finished output. long tasks are where the difference shows. try it on open source coworker Eigent. 🦾

Eigent AI

22,881 views • 1 month ago

I tested GLM 5.2 on three real builds. A horror house game. A 3D stealth game. A Remotion marketing video. The results are exactly what I expect from Chinese open source coding models right now. The horror house has broken game logic. Collect all three keys, complete the objective, and the front door still won't open. The stealth game shipped with lighting so dark you can't see your own character. The Remotion video is rough. GLM 5.2 is faster than 5.1 and 10x cheaper than the frontier. But Opus 4.8 and GPT 5.5 are still in a different league for serious coding work. Chinese models keep missing the attention to detail that matters. I MISS FABLE 5 ALREADY. Full review live now.

I tested GLM 5.2 on three real builds. A horror house game. A 3D stealth game. A Remotion marketing video. The results are exactly what I expect from Chinese open source coding models right now. The horror house has broken game logic. Collect all three keys, complete the objective, and the front door still won't open. The stealth game shipped with lighting so dark you can't see your own character. The Remotion video is rough. GLM 5.2 is faster than 5.1 and 10x cheaper than the frontier. But Opus 4.8 and GPT 5.5 are still in a different league for serious coding work. Chinese models keep missing the attention to detail that matters. I MISS FABLE 5 ALREADY. Full review live now.

BridgeMind

24,527 views • 1 month ago

🎉 Day-0 support for in vLLM, available today in v0.23.0! Congrats to Z.ai on GLM-5.2, a flagship model built for long-horizon coding agents. ✨ 1M-token context, built to hold project-scale engineering work in a single run ✨ Tuned for long-horizon coding: large-scale implementation, automated research, and performance optimization ✨ One task can carry a full dev workflow, from requirements to a deployable product across platforms ✨ Client-side and mobile engineering, including an on-device debugging loop Try it out running it on vLLM today: 🔗

🎉 Day-0 support for in vLLM, available today in v0.23.0! Congrats to Z.ai on GLM-5.2, a flagship model built for long-horizon coding agents. ✨ 1M-token context, built to hold project-scale engineering work in a single run ✨ Tuned for long-horizon coding: large-scale implementation, automated research, and performance optimization ✨ One task can carry a full dev workflow, from requirements to a deployable product across platforms ✨ Client-side and mobile engineering, including an on-device debugging loop Try it out running it on vLLM today: 🔗

vLLM

36,019 views • 1 month ago

FABLE 5 JUST REBUILT FORTNITE FROM 1 PROMPT IN 20 MINUTES. EPIC NEEDED HUNDREDS OF ENGINEERS AND 6 YEARS. The test was simple: Fable 5 vs GLM 5.2, one prompt each, a 100-player battle royale running in the browser with no engine and no team. GLM went first and checked all 8 milestones. The game loaded with a battle bus, a generated map, Tilted Towers, even health and shield bars ripped straight from 2018. Then it fell apart. 2 FPS on landing, S walking forward and W walking backward, a gun that fires upward and a camera that refuses to look up. First fight ended with the enemy standing above physically impossible to aim at. Then Fable 5 took the same prompt on max effort. Different game entirely. Real character models instead of beans, hit reactions lighting up on contact, a shadow tracking every player, movement that just works out of the box. One follow-up prompt “add more Fortnite elements” and it shipped gliders, chests, shield potions, medkits and a sniper without breaking a thing. Final circle at 3:28 AM. Storm closing, 0 bullets left, box to box with medkit spam and a last-second swap to the SMG. Victory Royale in a game an AI wrote from a single sentence. GLM checked more boxes. Fable built something you could actually play and no benchmark on earth measures that gap. Epic needed 6 years. Fable 5 needed 1 prompt.

FABLE 5 JUST REBUILT FORTNITE FROM 1 PROMPT IN 20 MINUTES. EPIC NEEDED HUNDREDS OF ENGINEERS AND 6 YEARS. The test was simple: Fable 5 vs GLM 5.2, one prompt each, a 100-player battle royale running in the browser with no engine and no team. GLM went first and checked all 8 milestones. The game loaded with a battle bus, a generated map, Tilted Towers, even health and shield bars ripped straight from 2018. Then it fell apart. 2 FPS on landing, S walking forward and W walking backward, a gun that fires upward and a camera that refuses to look up. First fight ended with the enemy standing above physically impossible to aim at. Then Fable 5 took the same prompt on max effort. Different game entirely. Real character models instead of beans, hit reactions lighting up on contact, a shadow tracking every player, movement that just works out of the box. One follow-up prompt “add more Fortnite elements” and it shipped gliders, chests, shield potions, medkits and a sniper without breaking a thing. Final circle at 3:28 AM. Storm closing, 0 bullets left, box to box with medkit spam and a last-second swap to the SMG. Victory Royale in a game an AI wrote from a single sentence. GLM checked more boxes. Fable built something you could actually play and no benchmark on earth measures that gap. Epic needed 6 years. Fable 5 needed 1 prompt.

West Lord

80,912 views • 1 month ago

Excited to present the LLM-Council skill. Initial idea by Karpathy. I just packaged it as a skill. You can easily spin up a council of LLMs or agents via Fireworks. Watch how the new GLM-5 model "deliberates" on other LLMs' thoughts on the big question, "Can LLMs reason?" Things worth paying attention to: New open models like GLM-5 have surprisingly improved on complex reasoning and long-running agentic tasks. The AskUserQuestion tool in Claude Code came in handy to select the council and chairperson. As Andrej Karpathy puts it, it's a really interesting way to get different perspectives from LLMs, which can lead to better decision-making on whatever task you are working on. You can use it for other agentic coding use cases, like evaluation, tool building, designing, and research.

Excited to present the LLM-Council skill. Initial idea by Karpathy. I just packaged it as a skill. You can easily spin up a council of LLMs or agents via Fireworks. Watch how the new GLM-5 model "deliberates" on other LLMs' thoughts on the big question, "Can LLMs reason?" Things worth paying attention to: New open models like GLM-5 have surprisingly improved on complex reasoning and long-running agentic tasks. The AskUserQuestion tool in Claude Code came in handy to select the council and chairperson. As Andrej Karpathy puts it, it's a really interesting way to get different perspectives from LLMs, which can lead to better decision-making on whatever task you are working on. You can use it for other agentic coding use cases, like evaluation, tool building, designing, and research.

elvis

39,593 views • 5 months ago

Tencent just dropped Hy3, and it's worth a look if you're building AI agents. I spent some time putting it through its paces today. A quick rundown of what makes this release notable: → 295B total parameters (Mixture-of-Experts), but only 21B active at inference — a real efficiency play → 256K context window for handling large codebases or long documents → Purpose-built improvements for coding and multi-step agent tasks → Fully open under Apache 2.0 — no restrictive commercial terms → A free two-week API window currently live on OpenRouter On the practical side: prompts involving layered instructions and coding tasks came back fast and coherent. Tencent says this release builds directly on feedback from 50+ internal product teams following an earlier preview, with reported drops in hallucination rate (12.5% → 5.4%) and commonsense errors (25.4% → 12.7%). Those are Tencent's own numbers, so treat them as a starting point rather than gospel until third parties weigh in. For context: Tencent's own comparisons put Hy3 behind GLM-5.2 specifically on coding benchmarks — GLM is a much bigger model (~744B total), so the trade-off makes sense. Hy3's pitch isn't "biggest," it's "efficient enough to actually deploy." Bottom line — if you're evaluating open-weight options for agent workloads, this is a solid one to add to the testing queue while the free window is open. Try it here: Tencent Hy #Hy3 #Hunyuan #TencentAI #AICoding

Tencent just dropped Hy3, and it's worth a look if you're building AI agents. I spent some time putting it through its paces today. A quick rundown of what makes this release notable: → 295B total parameters (Mixture-of-Experts), but only 21B active at inference — a real efficiency play → 256K context window for handling large codebases or long documents → Purpose-built improvements for coding and multi-step agent tasks → Fully open under Apache 2.0 — no restrictive commercial terms → A free two-week API window currently live on OpenRouter On the practical side: prompts involving layered instructions and coding tasks came back fast and coherent. Tencent says this release builds directly on feedback from 50+ internal product teams following an earlier preview, with reported drops in hallucination rate (12.5% → 5.4%) and commonsense errors (25.4% → 12.7%). Those are Tencent's own numbers, so treat them as a starting point rather than gospel until third parties weigh in. For context: Tencent's own comparisons put Hy3 behind GLM-5.2 specifically on coding benchmarks — GLM is a much bigger model (~744B total), so the trade-off makes sense. Hy3's pitch isn't "biggest," it's "efficient enough to actually deploy." Bottom line — if you're evaluating open-weight options for agent workloads, this is a solid one to add to the testing queue while the free window is open. Try it here: Tencent Hy #Hy3 #Hunyuan #TencentAI #AICoding

Felix

36,925 views • 28 days ago

sonnet 5 vs sonnet 4.6 vs opus 4.8 vs glm 5.2 – frontend tasks dropped sonnet 5 into a quick test today. same three prompts to all four models, single-shot html/canvas, no edits: • objects falling on a trampoline • rockets playing tennis • a slingshot breaking bottles ranked by speed (total across the 3 tasks): 1. opus 4.8 – 15m 09s 2. sonnet 5 – 16m 05s 3. glm 5.2 – 27m 18s 4. sonnet 4.6 – 35m 06s ranked by code shortness (total loc): 1. sonnet 5 – 1794 2. opus 4.8 – 2063 3. sonnet 4.6 – 2182 4. glm 5.2 – 3285 sonnet 5 came out on top here – leanest code overall and a near-tie for fastest it was also the most creative. in every task it added something none of the others did: – kept the trampoline vibrating after the objects landed – drew a +1 next to the rocket that scored the point – turned the slingshot to face the next bottle before each shot opus 4.8 evaluated the code sonnet 5 produced. four things stood out: • the sphere is a fake, and that's the smart move. the cube and star are real 3d meshes with proper culling and shading, but the ball is just a flat shaded circle. a lit sphere looks identical from every angle, so building it in 3d would burn compute for zero visible payoff. knowing where not to bother is its own kind of skill • weight actually means something on the trampoline. the star is heavy, so it barely bounces and dents the mat hard. the ball is light, so it's lively and leaves a shallow dip. the three objects aren't just different shapes – they have different temperaments, and the physics is what gives them that • the slingshot is framed like a shot, not just drawn. the handle is anchored below the bottom of the screen and runs off-frame, so it reads as something you're holding rather than a sprite parked in the scene. that's a staging instinct, not a rendering one • the paddle ai forward-simulates the ball to predict where it'll land, then adds a deliberate error bias (roughly 1 in 5 shots is a real miss). that's why scoring looks natural instead of robotic – plus four distinct fault types with a catch-all so a rally never hangs without a result bottom line: sonnet 5 does more with less. fastest tier, leanest code, and the only one that added small touches nobody asked for follow thehype. for 24/7 ai news, analysis and breakdowns

sonnet 5 vs sonnet 4.6 vs opus 4.8 vs glm 5.2 – frontend tasks dropped sonnet 5 into a quick test today. same three prompts to all four models, single-shot html/canvas, no edits: • objects falling on a trampoline • rockets playing tennis • a slingshot breaking bottles ranked by speed (total across the 3 tasks): 1. opus 4.8 – 15m 09s 2. sonnet 5 – 16m 05s 3. glm 5.2 – 27m 18s 4. sonnet 4.6 – 35m 06s ranked by code shortness (total loc): 1. sonnet 5 – 1794 2. opus 4.8 – 2063 3. sonnet 4.6 – 2182 4. glm 5.2 – 3285 sonnet 5 came out on top here – leanest code overall and a near-tie for fastest it was also the most creative. in every task it added something none of the others did: – kept the trampoline vibrating after the objects landed – drew a +1 next to the rocket that scored the point – turned the slingshot to face the next bottle before each shot opus 4.8 evaluated the code sonnet 5 produced. four things stood out: • the sphere is a fake, and that's the smart move. the cube and star are real 3d meshes with proper culling and shading, but the ball is just a flat shaded circle. a lit sphere looks identical from every angle, so building it in 3d would burn compute for zero visible payoff. knowing where not to bother is its own kind of skill • weight actually means something on the trampoline. the star is heavy, so it barely bounces and dents the mat hard. the ball is light, so it's lively and leaves a shallow dip. the three objects aren't just different shapes – they have different temperaments, and the physics is what gives them that • the slingshot is framed like a shot, not just drawn. the handle is anchored below the bottom of the screen and runs off-frame, so it reads as something you're holding rather than a sprite parked in the scene. that's a staging instinct, not a rendering one • the paddle ai forward-simulates the ball to predict where it'll land, then adds a deliberate error bias (roughly 1 in 5 shots is a real miss). that's why scoring looks natural instead of robotic – plus four distinct fault types with a catch-all so a rally never hangs without a result bottom line: sonnet 5 does more with less. fastest tier, leanest code, and the only one that added small touches nobody asked for follow thehype. for 24/7 ai news, analysis and breakdowns

thehype.

14,518 views • 1 month ago

Crazy... I just found a super cheap way to code with AI Tried GLM 4.5 on Claude Code and I’m really impressed! Prompt: “create an interactive python app where users can track, manage or analyze data” And, It came up with: • Dashboard - a clean interface to see all your data at a glance • Datasets - manage multiple collections of data easily • Data entries - add, edit, and organize your records without hassle • Analysis - get insights and summaries directly from your data What I observed: Everything worked smoothly, GLM 4.5 is definitely one of the best open-source models for coding, I’ll keep using it. Check it out now: And the best part: super affordable coding plans! • Lite - $3/month (promo). ~120 prompts every 5 hours. Works with Claude Code, Roo Code, Cline, Kilo Code, and more. • Pro - $15/month (promo). 5× more usage than Lite. Faster. Includes image/video tools and web search. • Max - $30/month (promo). 4× more usage than Pro. Guaranteed performance. Early access to new features. Bonus - Use my referral link to get an extra 10% off on top of these promo prices! Video demo 👇

Crazy... I just found a super cheap way to code with AI Tried GLM 4.5 on Claude Code and I’m really impressed! Prompt: “create an interactive python app where users can track, manage or analyze data” And, It came up with: • Dashboard - a clean interface to see all your data at a glance • Datasets - manage multiple collections of data easily • Data entries - add, edit, and organize your records without hassle • Analysis - get insights and summaries directly from your data What I observed: Everything worked smoothly, GLM 4.5 is definitely one of the best open-source models for coding, I’ll keep using it. Check it out now: And the best part: super affordable coding plans! • Lite - $3/month (promo). ~120 prompts every 5 hours. Works with Claude Code, Roo Code, Cline, Kilo Code, and more. • Pro - $15/month (promo). 5× more usage than Lite. Faster. Includes image/video tools and web search. • Max - $30/month (promo). 4× more usage than Pro. Guaranteed performance. Early access to new features. Bonus - Use my referral link to get an extra 10% off on top of these promo prices! Video demo 👇

Chidanand Tripathi

83,854 views • 10 months ago

People made fun of Alex Finn for buying three Mac Studios to run AI at home. Then Fable got banned for a week, GLM 5.2 dropped, and those exact Mac Studios started reselling for 4x what he paid. He showed me how he built his home AI lab from scratch. Here's the playbook: 1) The hardware. three 512GB Mac Studios, an NVIDIA DGX Spark, a custom RTX 5090 build, and a few Mac Minis. ~$30k all in. 2) The buying framework... - Mac Studio: huge memory, runs GLM 5.2 (open weights, near Opus 4.8 on benchmarks), but slow. - DGX Spark ($4,800): the sweet spot for most people. - RTX 5090: smaller models at blazing speed (Qwen's 29B now hits Sonnet 4 level). 3) Tailscale networks every machine into one private network with root access to each other. Only one machine is plugged into a monitor. 4) A Nous Research Hermes agent is his IT guy. New model drops? It SSHs into the right box, loads 5 candidates, runs evals overnight, and reports back which task belongs on which machine. Alex has literally never loaded a model himself. 5) The whole point: achieving "ambient intelligence." Always-on jobs that would bankrupt you on per-token billing. A security sweep of his API endpoints every hour. Code optimization every 20 minutes. Database anomaly & churn detection. Hourly scraping of X, Reddit & Hacker News for business opportunities. 6) Running those workloads on frontier models would cost thousands a month. His actual cost: ~$60 more in electricity. 7) Btw he's not anti-frontier. He still maxes out his Claude plan. The way he sees it: frontier is for hard thinking, local is for the foot soldiers that never sleep. 8) "We own everything except for the intelligence. Why can't we own the intelligence?" 9) He thinks frontier-level intelligence runs on consumer hardware within 6 months.

People made fun of Alex Finn for buying three Mac Studios to run AI at home. Then Fable got banned for a week, GLM 5.2 dropped, and those exact Mac Studios started reselling for 4x what he paid. He showed me how he built his home AI lab from scratch. Here's the playbook: 1) The hardware. three 512GB Mac Studios, an NVIDIA DGX Spark, a custom RTX 5090 build, and a few Mac Minis. ~$30k all in. 2) The buying framework... - Mac Studio: huge memory, runs GLM 5.2 (open weights, near Opus 4.8 on benchmarks), but slow. - DGX Spark ($4,800): the sweet spot for most people. - RTX 5090: smaller models at blazing speed (Qwen's 29B now hits Sonnet 4 level). 3) Tailscale networks every machine into one private network with root access to each other. Only one machine is plugged into a monitor. 4) A Nous Research Hermes agent is his IT guy. New model drops? It SSHs into the right box, loads 5 candidates, runs evals overnight, and reports back which task belongs on which machine. Alex has literally never loaded a model himself. 5) The whole point: achieving "ambient intelligence." Always-on jobs that would bankrupt you on per-token billing. A security sweep of his API endpoints every hour. Code optimization every 20 minutes. Database anomaly & churn detection. Hourly scraping of X, Reddit & Hacker News for business opportunities. 6) Running those workloads on frontier models would cost thousands a month. His actual cost: ~$60 more in electricity. 7) Btw he's not anti-frontier. He still maxes out his Claude plan. The way he sees it: frontier is for hard thinking, local is for the foot soldiers that never sleep. 8) "We own everything except for the intelligence. Why can't we own the intelligence?" 9) He thinks frontier-level intelligence runs on consumer hardware within 6 months.

Alex Lieberman

57,688 views • 28 days ago

Six open-source LLMs. One sliding puzzle. A brutal test of long-horizon reasoning and tool calling. Five of them broke. One didn't. I gave each model a move_tile tool and a scrambled 3×3 board, then asked it to solve the puzzle through pure turn-by-turn reasoning. The deeper the scramble, the more brutal the search. Five runs per depth, best run kept. A model fails the round if it exceeds 6x the optimal move count. > Depth 5: Everyone solves it. Yawn. > Depth 10: GLM 5.1 melts down. 43 moves. Cut. > Depth 12: Gemma4 26B loses the plot, shuffling tiles in circles. Gone. > Depth 15: The wall. DeepSeek V4 Flash, out. DeepSeek V4 Pro, out. Gemma4, out again. GLM 5.1, out. Two survivors: Qwen3.6 35B-A3B, and Kimi K2.6 with an 11-move solve that looked like cheating. > Depth 18: Same two. Everyone else hallucinating tiles that weren't there. > Depth 22: Final boss. Kimi, flawless for five rounds, finally cracks. 81 moves. Still scrambled. DeepSeek V4 Pro limps home at 90. Qwen3.6 35B-A3B solves it in 36. The smallest model in the room. ~3B active params. Fits on a single 3090. It beat everything. Kimi was elegant. Qwen3.6 was unstoppable.

Six open-source LLMs. One sliding puzzle. A brutal test of long-horizon reasoning and tool calling. Five of them broke. One didn't. I gave each model a move_tile tool and a scrambled 3×3 board, then asked it to solve the puzzle through pure turn-by-turn reasoning. The deeper the scramble, the more brutal the search. Five runs per depth, best run kept. A model fails the round if it exceeds 6x the optimal move count. > Depth 5: Everyone solves it. Yawn. > Depth 10: GLM 5.1 melts down. 43 moves. Cut. > Depth 12: Gemma4 26B loses the plot, shuffling tiles in circles. Gone. > Depth 15: The wall. DeepSeek V4 Flash, out. DeepSeek V4 Pro, out. Gemma4, out again. GLM 5.1, out. Two survivors: Qwen3.6 35B-A3B, and Kimi K2.6 with an 11-move solve that looked like cheating. > Depth 18: Same two. Everyone else hallucinating tiles that weren't there. > Depth 22: Final boss. Kimi, flawless for five rounds, finally cracks. 81 moves. Still scrambled. DeepSeek V4 Pro limps home at 90. Qwen3.6 35B-A3B solves it in 36. The smallest model in the room. ~3B active params. Fits on a single 3090. It beat everything. Kimi was elegant. Qwen3.6 was unstoppable.

stevibe

16,434 views • 2 months ago

How to setup a multi agent system? Bookmark it 📂 "The Trading Floor" Multi-Agent Market Analysis Council to analyze a stock ticker Z.ai GLM-4.7 🤝 OpenCode Agent framework: CrewAI How it works? 1. User enters a stock ticker to analyze 2. 5 AI agents wake up, each with distinct expertise: - Quant Analyst — technical indicators & price patterns - Sentiment Scout — market mood & crowd psychology - Macro Strategist — sector dynamics & economic context - Risk Manager — volatility, drawdowns & position sizing - Portfolio Chief — synthesizes all perspectives 3. Agents analyze independently using real market data 4. They debate, challenge assumptions, and identify disagreements 5. Portfolio Chief resolves conflicts and delivers a consensus recommendation 6. Final output: buy/hold/sell rating with confidence level, position size, and key risks How to built The Trading Floor? 1. Chose CrewAI as the agent framework — handles multi-agent orchestration out of the box 2. Defined 5 agents with distinct roles, goals, and backstories in Python 3. Built custom tools wrapping yfinance for real market data (prices, indicators, volatility) 4. Configured sequential workflow — specialists analyze first, Portfolio Chief synthesizes last 5. Set up FastAPI backend with SSE to stream agent thoughts in real-time 6. Built Next.js frontend to visualize the "board of directors" deliberating live 7. One environment variable (MODEL=openai/gpt-5.2) powers all agents 8. Generated unique agent icons with AI image tools Total cost: $0 for the framework, pay only for LLM API calls Tech stack: - GLM-4.7 with opencode to build the app - CrewAI (open source) for agent orchestration - GPT-5.2 powering each agent - FastAPI + SSE for real-time streaming - Next.js frontend showing live agent deliberations

How to setup a multi agent system? Bookmark it 📂 "The Trading Floor" Multi-Agent Market Analysis Council to analyze a stock ticker Z.ai GLM-4.7 🤝 OpenCode Agent framework: CrewAI How it works? 1. User enters a stock ticker to analyze 2. 5 AI agents wake up, each with distinct expertise: - Quant Analyst — technical indicators & price patterns - Sentiment Scout — market mood & crowd psychology - Macro Strategist — sector dynamics & economic context - Risk Manager — volatility, drawdowns & position sizing - Portfolio Chief — synthesizes all perspectives 3. Agents analyze independently using real market data 4. They debate, challenge assumptions, and identify disagreements 5. Portfolio Chief resolves conflicts and delivers a consensus recommendation 6. Final output: buy/hold/sell rating with confidence level, position size, and key risks How to built The Trading Floor? 1. Chose CrewAI as the agent framework — handles multi-agent orchestration out of the box 2. Defined 5 agents with distinct roles, goals, and backstories in Python 3. Built custom tools wrapping yfinance for real market data (prices, indicators, volatility) 4. Configured sequential workflow — specialists analyze first, Portfolio Chief synthesizes last 5. Set up FastAPI backend with SSE to stream agent thoughts in real-time 6. Built Next.js frontend to visualize the "board of directors" deliberating live 7. One environment variable (MODEL=openai/gpt-5.2) powers all agents 8. Generated unique agent icons with AI image tools Total cost: $0 for the framework, pay only for LLM API calls Tech stack: - GLM-4.7 with opencode to build the app - CrewAI (open source) for agent orchestration - GPT-5.2 powering each agent - FastAPI + SSE for real-time streaming - Next.js frontend showing live agent deliberations

CloudAI-X

57,703 views • 7 months ago

AgentRouter is handing out $125 in FREE API credits for OPUS 4.8 > No card. > GitHub sign-in. > One API key for Opus 4.8 / Opus 4.7 / Sonnet 4.6 / GLM 5.1 The useful part is not “another model gateway.” It is that you can point coding tools and agents at one OpenAI-compatible base URL instead of juggling separate keys for Claude and GLM, and others. What you get: • $125 signup balance • one API key • OpenAI-compatible endpoints • chat, responses, messages, embeddings, images, audio, rerank endpoints listed • claimed support for Cursor, Claude Code, Hermes, and other agent tools The flow is simple: 1. go to AgentRouter (link in comment) 2. sign in or create an account 3. generate an API key in the console 4. copy the base URL 5. paste it into Cursor / Claude Code / your agent runtime 6. run a small test before moving anything serious Caveat: AgentRouter’s own site currently has a notice saying Claude-series service was recently hit by stability issues and large-scale Claude access was suspended. So don’t treat this as “guaranteed free Claude forever.” Treat it as $125 of free routing credits to test which models are live, how the latency feels, and whether the gateway is stable enough for your workflow. Still a very solid save if you build with agents. Most people burn paid API credits just testing configs. This gives you room to test first, then decide if it belongs in your stack.

AgentRouter is handing out $125 in FREE API credits for OPUS 4.8 > No card. > GitHub sign-in. > One API key for Opus 4.8 / Opus 4.7 / Sonnet 4.6 / GLM 5.1 The useful part is not “another model gateway.” It is that you can point coding tools and agents at one OpenAI-compatible base URL instead of juggling separate keys for Claude and GLM, and others. What you get: • $125 signup balance • one API key • OpenAI-compatible endpoints • chat, responses, messages, embeddings, images, audio, rerank endpoints listed • claimed support for Cursor, Claude Code, Hermes, and other agent tools The flow is simple: 1. go to AgentRouter (link in comment) 2. sign in or create an account 3. generate an API key in the console 4. copy the base URL 5. paste it into Cursor / Claude Code / your agent runtime 6. run a small test before moving anything serious Caveat: AgentRouter’s own site currently has a notice saying Claude-series service was recently hit by stability issues and large-scale Claude access was suspended. So don’t treat this as “guaranteed free Claude forever.” Treat it as $125 of free routing credits to test which models are live, how the latency feels, and whether the gateway is stable enough for your workflow. Still a very solid save if you build with agents. Most people burn paid API credits just testing configs. This gives you room to test first, then decide if it belongs in your stack.

kocer

24,905 views • 1 month ago

JARVIS-1: Open-World Multi-task Agents with Memory-Augmented Multimodal Language Models paper page: Achieving human-like planning and control with multimodal observations in an open world is a key milestone for more functional generalist agents. Existing approaches can handle certain long-horizon tasks in an open world. However, they still struggle when the number of open-world tasks could potentially be infinite and lack the capability to progressively enhance task completion as game time progresses. We introduce JARVIS-1, an open-world agent that can perceive multimodal input (visual observations and human instructions), generate sophisticated plans, and perform embodied control, all within the popular yet challenging open-world Minecraft universe. Specifically, we develop JARVIS-1 on top of pre-trained multimodal language models, which map visual observations and textual instructions to plans. The plans will be ultimately dispatched to the goal-conditioned controllers. We outfit JARVIS-1 with a multimodal memory, which facilitates planning using both pre-trained knowledge and its actual game survival experiences. In our experiments, JARVIS-1 exhibits nearly perfect performances across over 200 varying tasks from the Minecraft Universe Benchmark, ranging from entry to intermediate levels. JARVIS-1 has achieved a completion rate of 12.5% in the long-horizon diamond pickaxe task. This represents a significant increase up to 5 times compared to previous records. Furthermore, we show that JARVIS-1 is able to self-improve following a life-long learning paradigm thanks to multimodal memory, sparking a more general intelligence and improved autonomy.

JARVIS-1: Open-World Multi-task Agents with Memory-Augmented Multimodal Language Models paper page: Achieving human-like planning and control with multimodal observations in an open world is a key milestone for more functional generalist agents. Existing approaches can handle certain long-horizon tasks in an open world. However, they still struggle when the number of open-world tasks could potentially be infinite and lack the capability to progressively enhance task completion as game time progresses. We introduce JARVIS-1, an open-world agent that can perceive multimodal input (visual observations and human instructions), generate sophisticated plans, and perform embodied control, all within the popular yet challenging open-world Minecraft universe. Specifically, we develop JARVIS-1 on top of pre-trained multimodal language models, which map visual observations and textual instructions to plans. The plans will be ultimately dispatched to the goal-conditioned controllers. We outfit JARVIS-1 with a multimodal memory, which facilitates planning using both pre-trained knowledge and its actual game survival experiences. In our experiments, JARVIS-1 exhibits nearly perfect performances across over 200 varying tasks from the Minecraft Universe Benchmark, ranging from entry to intermediate levels. JARVIS-1 has achieved a completion rate of 12.5% in the long-horizon diamond pickaxe task. This represents a significant increase up to 5 times compared to previous records. Furthermore, we show that JARVIS-1 is able to self-improve following a life-long learning paradigm thanks to multimodal memory, sparking a more general intelligence and improved autonomy.

AK

141,440 views • 2 years ago

Amp CEO, Quinn Slack on why he thinks Kimi K3 could be a Sputnik moment for America: "With GLM 5.2 we had a mini Sputnik moment. When the Soviet Union launched a satellite into space before the US and we realized we have to catch up." "Kimi K3 I need to go and see, but that could be an even bigger Sputnik moment that wakes us up and will probably create a lot of good effects in terms of catalyzing the understanding of AI as a strategic advantage for the US rather than as something people should be against." "But I think it could also scare a lot of people, and I can understand why they'd be scared." "The worst possible outcome could be that open models are banned or are banned from certain industries or companies in the US. There's a lot of people in lobbying organizations in Washington DC, and a lot of people in various parts of the government that want to do that." "They want to say, we won't let these models trained in China be used in the US. There's a lot of people trying to do that now, and that would make us all worse off. I'm pretty nervous about that." Quinn Slack Amp

Amp CEO, Quinn Slack on why he thinks Kimi K3 could be a Sputnik moment for America: "With GLM 5.2 we had a mini Sputnik moment. When the Soviet Union launched a satellite into space before the US and we realized we have to catch up." "Kimi K3 I need to go and see, but that could be an even bigger Sputnik moment that wakes us up and will probably create a lot of good effects in terms of catalyzing the understanding of AI as a strategic advantage for the US rather than as something people should be against." "But I think it could also scare a lot of people, and I can understand why they'd be scared." "The worst possible outcome could be that open models are banned or are banned from certain industries or companies in the US. There's a lot of people in lobbying organizations in Washington DC, and a lot of people in various parts of the government that want to do that." "They want to say, we won't let these models trained in China be used in the US. There's a lot of people trying to do that now, and that would make us all worse off. I'm pretty nervous about that." Quinn Slack Amp

MTS

58,660 views • 19 days ago

timelapse #160 (13.5 hrs) - grok 4.5 high in grok build was my main for today. just getting a vibe check. it understands my intent really well and never messed anything up. i do prefer to stay in the loop with it more than claude though. SpaceXAI really cooked! - not sure what it was, but i was able to stay in flow state with grok build for hours and hours. my memory and retention helped me get a lot done - exhausted 4th claude max 20x plan in 48 hours - training dspark drafters on my traces (kept gpu warm the whole time) - did grok 4.5 high sweeps and was quite impressed at how far the model has come. its around gpt 5.5 level (official post coming) - some very important meetings - contract work - optimizing amd kernels for glm 5.2 fp8 (mostly custom all-reduce whilst avoid deadlocks) for the tiny corp - one shotted and ran all frontier models on a new benchmark which ill be revealing shortly. its the toughest benchmark ive ever made. fable doesnt even come close to beating it - had grok 4.5 take over my minecraft C/CUDA implementation. felt like whack a mole but maybe i just wasnt using it right

timelapse #160 (13.5 hrs) - grok 4.5 high in grok build was my main for today. just getting a vibe check. it understands my intent really well and never messed anything up. i do prefer to stay in the loop with it more than claude though. SpaceXAI really cooked! - not sure what it was, but i was able to stay in flow state with grok build for hours and hours. my memory and retention helped me get a lot done - exhausted 4th claude max 20x plan in 48 hours - training dspark drafters on my traces (kept gpu warm the whole time) - did grok 4.5 high sweeps and was quite impressed at how far the model has come. its around gpt 5.5 level (official post coming) - some very important meetings - contract work - optimizing amd kernels for glm 5.2 fp8 (mostly custom all-reduce whilst avoid deadlocks) for the tiny corp - one shotted and ran all frontier models on a new benchmark which ill be revealing shortly. its the toughest benchmark ive ever made. fable doesnt even come close to beating it - had grok 4.5 take over my minecraft C/CUDA implementation. felt like whack a mole but maybe i just wasnt using it right

Elliot Arledge

13,619 views • 26 days ago

introducing OpenBench v1, an open framework for measuring AI performance and efficiency for your codebase and use case. Companies are realizing that you can't simply tokenmaxx, and are frantically looking for better ways to use and measure AI use and efficiency. One approach is evidenced by the rise in model routers: OpenRouter, Cloudflare, Databricks, Vercel, and now Ramp to name a few. But they'll also need the ability to evaluate how agents are performing in their actual use cases and codebase. A good example is DoorDash's recent evals on their codebase and prs. This will become even more important as companies explore different model + harness combinations. OpenBench will be built to this direction, focusing on the cross between correctness vs efficiency in both token use and latency. Starting off, OpenBench makes it easy for anyone to add to the task set, and run a variety of harness + models. The framework also makes it easy to add any custom harness variant, for example I've been testing codex variants with ablations against the stock codex harness. OpenBench will also have tooling to help with discovering/adding reliable tasks from your repo. Today I have integrated harnesses: codex, claude, cursor, devin, grok build, pi, and eval'ed them for different models like gpt 5.6-sol, GLM 5.2, Kimi K3, Grok 4.5, etc. measuring correctness %, token use (in/out/cache), and latency.

introducing OpenBench v1, an open framework for measuring AI performance and efficiency for your codebase and use case. Companies are realizing that you can't simply tokenmaxx, and are frantically looking for better ways to use and measure AI use and efficiency. One approach is evidenced by the rise in model routers: OpenRouter, Cloudflare, Databricks, Vercel, and now Ramp to name a few. But they'll also need the ability to evaluate how agents are performing in their actual use cases and codebase. A good example is DoorDash's recent evals on their codebase and prs. This will become even more important as companies explore different model + harness combinations. OpenBench will be built to this direction, focusing on the cross between correctness vs efficiency in both token use and latency. Starting off, OpenBench makes it easy for anyone to add to the task set, and run a variety of harness + models. The framework also makes it easy to add any custom harness variant, for example I've been testing codex variants with ablations against the stock codex harness. OpenBench will also have tooling to help with discovering/adding reliable tasks from your repo. Today I have integrated harnesses: codex, claude, cursor, devin, grok build, pi, and eval'ed them for different models like gpt 5.6-sol, GLM 5.2, Kimi K3, Grok 4.5, etc. measuring correctness %, token use (in/out/cache), and latency.

Matthew Lam

120,732 views • 14 days ago

Which LLM reasons best when it doesn't have all the information? Enter LLM Poker Arena to find out. It's a Poker Playing benchmark where top reasoning models play Texas Hold'em poker against each other. Claude Opus 4.5, GPT-5.2, Gemini 2.5 Pro, and Grok 4 all sit at the same table and play full tournaments to see who finishes with the chips. Poker is very different when it comes to reasoning. It has to balance probabilistic reasoning, opponent modeling and make decisions under uncertainty. Poker is an interesting evaluation because it tests reasoning under incomplete information, something most coding benchmarks do not capture. In this tournaments the rules are: - Each LLM starts with $1,000 chips - Small and big blinds start at $25 / $50 - Blinds double every 3 minutes - All models run in their reasoning or thinking modes After the first 5 tournaments: - Claude Opus 4.5 with Thinking has 3 wins - GPT-5.2 has 2 wins - Grok 4 and Gemini 2.5 Pro have 0 wins Early results suggest Claude performs quite well at poker as well. Also five is a very small sample size. Planning to run many more tournaments, publish the full benchmark data and add a prediction market on top of it. Thanks for the suggestion clipz. Much more coming as part of Poker Cities !! This was built on Replit ⠕ using their AI integrations, which made it straightforward to connect Claude, GPT, and Gemini. What model do you think wins after 100 tournaments?

Which LLM reasons best when it doesn't have all the information? Enter LLM Poker Arena to find out. It's a Poker Playing benchmark where top reasoning models play Texas Hold'em poker against each other. Claude Opus 4.5, GPT-5.2, Gemini 2.5 Pro, and Grok 4 all sit at the same table and play full tournaments to see who finishes with the chips. Poker is very different when it comes to reasoning. It has to balance probabilistic reasoning, opponent modeling and make decisions under uncertainty. Poker is an interesting evaluation because it tests reasoning under incomplete information, something most coding benchmarks do not capture. In this tournaments the rules are: - Each LLM starts with $1,000 chips - Small and big blinds start at $25 / $50 - Blinds double every 3 minutes - All models run in their reasoning or thinking modes After the first 5 tournaments: - Claude Opus 4.5 with Thinking has 3 wins - GPT-5.2 has 2 wins - Grok 4 and Gemini 2.5 Pro have 0 wins Early results suggest Claude performs quite well at poker as well. Also five is a very small sample size. Planning to run many more tournaments, publish the full benchmark data and add a prediction market on top of it. Thanks for the suggestion clipz. Much more coming as part of Poker Cities !! This was built on Replit ⠕ using their AI integrations, which made it straightforward to connect Claude, GPT, and Gemini. What model do you think wins after 100 tournaments?

Anshul Dhawan

32,192 views • 6 months ago

This is the moment Chinese AI beat American AI. One of the largest public crypto companies in the world just DUMPED OpenAI and Anthropic. Coinbase switched to open-weight Chinese models from Zhipu and DeepSeek, and shaved nearly 50% off the company's internal AI spending. The numbers are absolutely ridiculous: Running the same enterprise workload through Anthropic's Claude costs $4,811. Running it through Zhipu's GLM 5.2 costs $544. That's a 9x price difference for equivalent output. OpenAI's GPT-5.5 sits in the middle at $3,357. DeepSeek's V4 lands at $1,071. Moonshot's Kimi at $948. On the actual benchmarks: Zhipu's GLM 5.2 scored 62.1 on SWE-bench Pro, the gold standard for coding. OpenAI's GPT-5.5 scored 58.6. One AI researcher called GLM 5.2 "at least as good as Opus 4.8 and GPT 5.5." Another called it "the first open model that can really compete with closed-source systems." The Chinese models are not just cheaper but they are now also beating American models on the benchmarks American companies pay $4,811 per workload for. Coinbase did the math first and reacted - more companies will certainly follow. Now watch what happens to the IPO timeline: Anthropic confidentially filed for an IPO targeting October at a $965 billion valuation. OpenAI followed days later with its own confidential filing. Both companies built their financial models on the assumption that they could keep charging enterprise prices that are 9 to 33x what Chinese competitors charge for the same task. Brian Armstrong publicly proved customers WILL leave. 45% of companies are now spending over $100,000 per month on AI, up from 20% last year. Every one of those customers is one quarterly budget review away from dumping American AI. OpenAI has reportedly already started preparing major token price cuts. Anthropic is expected to follow. And here's the thing... The export controls were supposed to CRUSH Chinese AI. The US government banned American AI chips, restricted model weights, blacklisted Alibaba and Baidu as Chinese military companies, and just banned Anthropic's flagship model from every foreign national on the planet. The entire premise of the American AI valuation bubble is that Washington can keep China two generations behind. But Chinese labs responded by building cheaper, more efficient models on inferior hardware and pricing them at one ninth the cost of the American alternative. And now American companies are voting with their checkbooks. The dominant American labs are valued at nearly $2 trillion combined on the assumption that their pricing power is durable. Coinbase proved it is not, and every customer doing a year-end budget review will be looking at the same math. For investors, the question here is what happens to the Anthropic IPO at $965 billion when the company is being forced to cut prices to defend share against open-weight Chinese models that score higher on the benchmarks. For everyone else, the bigger question is what happens when Washington spent four years and billions of dollars trying to contain Chinese AI, and the only thing that actually shifted in the end was American customers.

This is the moment Chinese AI beat American AI. One of the largest public crypto companies in the world just DUMPED OpenAI and Anthropic. Coinbase switched to open-weight Chinese models from Zhipu and DeepSeek, and shaved nearly 50% off the company's internal AI spending. The numbers are absolutely ridiculous: Running the same enterprise workload through Anthropic's Claude costs $4,811. Running it through Zhipu's GLM 5.2 costs $544. That's a 9x price difference for equivalent output. OpenAI's GPT-5.5 sits in the middle at $3,357. DeepSeek's V4 lands at $1,071. Moonshot's Kimi at $948. On the actual benchmarks: Zhipu's GLM 5.2 scored 62.1 on SWE-bench Pro, the gold standard for coding. OpenAI's GPT-5.5 scored 58.6. One AI researcher called GLM 5.2 "at least as good as Opus 4.8 and GPT 5.5." Another called it "the first open model that can really compete with closed-source systems." The Chinese models are not just cheaper but they are now also beating American models on the benchmarks American companies pay $4,811 per workload for. Coinbase did the math first and reacted - more companies will certainly follow. Now watch what happens to the IPO timeline: Anthropic confidentially filed for an IPO targeting October at a $965 billion valuation. OpenAI followed days later with its own confidential filing. Both companies built their financial models on the assumption that they could keep charging enterprise prices that are 9 to 33x what Chinese competitors charge for the same task. Brian Armstrong publicly proved customers WILL leave. 45% of companies are now spending over $100,000 per month on AI, up from 20% last year. Every one of those customers is one quarterly budget review away from dumping American AI. OpenAI has reportedly already started preparing major token price cuts. Anthropic is expected to follow. And here's the thing... The export controls were supposed to CRUSH Chinese AI. The US government banned American AI chips, restricted model weights, blacklisted Alibaba and Baidu as Chinese military companies, and just banned Anthropic's flagship model from every foreign national on the planet. The entire premise of the American AI valuation bubble is that Washington can keep China two generations behind. But Chinese labs responded by building cheaper, more efficient models on inferior hardware and pricing them at one ninth the cost of the American alternative. And now American companies are voting with their checkbooks. The dominant American labs are valued at nearly $2 trillion combined on the assumption that their pricing power is durable. Coinbase proved it is not, and every customer doing a year-end budget review will be looking at the same math. For investors, the question here is what happens to the Anthropic IPO at $965 billion when the company is being forced to cut prices to defend share against open-weight Chinese models that score higher on the benchmarks. For everyone else, the bigger question is what happens when Washington spent four years and billions of dollars trying to contain Chinese AI, and the only thing that actually shifted in the end was American customers.

Ricardo

251,902 views • 1 month ago

One of the things I’m most excited about this year is building agents that can work productively for hours, days, or weeks. Coding agents are starting to become very competent at this, but what about computer use agents? Our new benchmark, Odysseys (co-led with Lawrence Jang) is a set of 200 new tasks derived from real world browsing behavior that measure long horizon web navigation capabilities (potentially up to hours of web browsing work). Interestingly, we find that frontier CUAs are already surprisingly good at working productively for up to an hour on these tasks, but there’s a lot of work to be done in making them even more efficient. Like every other AI researcher, my real dream is to open a cafe once we solve ASI. So, here’s Opus 4.6 doing some market research for me ("I want to do market research on the most popular cafes in Singapore. Analyse the menus of the top 10 cafes in Singapore (by Google reviews/ratings), and make sure we include at least 1 from the North/South/East/West/Central regions of Singapore. Keep the relevant pages of each cafe open, and summarise their pricing, menu offerings, unique selling points, making sure to reference which tab is opened for each cafe. For each cafe, also help me figure out how long it would take to get to it from Tampines MRT, and include this in your final summary."). I was very impressed to see Opus 4.6 complete this task after working for 52 mins, satisfying all 7 rubrics that corresponded to this task. It provided a very nice markdown summary at the end that gave me all the information I asked for!

One of the things I’m most excited about this year is building agents that can work productively for hours, days, or weeks. Coding agents are starting to become very competent at this, but what about computer use agents? Our new benchmark, Odysseys (co-led with Lawrence Jang) is a set of 200 new tasks derived from real world browsing behavior that measure long horizon web navigation capabilities (potentially up to hours of web browsing work). Interestingly, we find that frontier CUAs are already surprisingly good at working productively for up to an hour on these tasks, but there’s a lot of work to be done in making them even more efficient. Like every other AI researcher, my real dream is to open a cafe once we solve ASI. So, here’s Opus 4.6 doing some market research for me ("I want to do market research on the most popular cafes in Singapore. Analyse the menus of the top 10 cafes in Singapore (by Google reviews/ratings), and make sure we include at least 1 from the North/South/East/West/Central regions of Singapore. Keep the relevant pages of each cafe open, and summarise their pricing, menu offerings, unique selling points, making sure to reference which tab is opened for each cafe. For each cafe, also help me figure out how long it would take to get to it from Tampines MRT, and include this in your final summary."). I was very impressed to see Opus 4.6 complete this task after working for 52 mins, satisfying all 7 rubrics that corresponded to this task. It provided a very nice markdown summary at the end that gave me all the information I asked for!

Jing Yu Koh

49,418 views • 3 months ago