Video wird geladen...

Video konnte nicht geladen werden

Zur Startseite

We've pushed a version update to the Terminal-Bench dataset and leaderboard. Terminal-Bench 4.0 calibrates task resources (time, cpu, memory), implements task fixes, and removes saturated tasks.

ryan_marten's profile picture

Ryan Marten

3,599 subscribers

107,624 Aufrufe • vor 2 Monaten •via X (Twitter)

Wissenschaft & Technologie
LIVE

Anya Rossi• Live Now

Private livecam show

0 Kommentare

Keine Kommentare verfügbar

Kommentare vom Original-Post werden hier angezeigt

Ähnliche Videos

Harbor integration is live with EvoSkill v.1.2.0 Harbor is a framework for evaluating AI agents against containerized benchmark tasks. It gives EvoSkill access to evolve agents against a registry of 190+ datasets — including benchmarks like SWE-bench Verified, Terminal-Bench 2.0, and Aider Polyglot. Here’s what it means for automated agent evolution ↓
0:16

Harbor integration is live with EvoSkill v.1.2.0 Harbor is a framework for evaluating AI agents against containerized benchmark tasks. It gives EvoSkill access to evolve agents against a registry of 190+ datasets — including benchmarks like SWE-bench Verified, Terminal-Bench 2.0, and Aider Polyglot. Here’s what it means for automated agent evolution ↓

Sentient

27,257 Aufrufe • vor 5 Monaten

If you want the task actually done, Grok 4.5 is making a very strong case. #1 on Long-Horizon Terminal-Bench. It completed 13 out of 46 long-horizon terminal tasks. The previous top models were sitting around 7. These aren't cute little prompts either. Multi-step workflows. Software engineering. Scientific computing. Millions of tokens. Long sessions. Grok 4.5 nearly doubled the mark. Turns out “get it done” is a very useful xAI feature. Grok xAI / Writer: Annette, Designer: Janné
0:16

If you want the task actually done, Grok 4.5 is making a very strong case. #1 on Long-Horizon Terminal-Bench. It completed 13 out of 46 long-horizon terminal tasks. The previous top models were sitting around 7. These aren't cute little prompts either. Multi-step workflows. Software engineering. Scientific computing. Millions of tokens. Long sessions. Grok 4.5 nearly doubled the mark. Turns out “get it done” is a very useful xAI feature. Grok xAI / Writer: Annette, Designer: Janné

Mario Nawfal

42,372 Aufrufe • vor 3 Monaten

Holy shit...Claude Code just killed one of the most annoying terminal problems — flickering. 🤯 Anthropic quietly shipped NO_FLICKER mode and it changes how AI feels inside the terminal: • No more screen flashing • Smooth scrolling + stable viewport • Mouse support (yes… in terminal) • Constant memory/CPU as chats grow • Cleaner selections (no UI junk copied) This isn’t just a UI tweak. They basically virtualized the entire terminal viewport to control rendering themselves. Result? Terminal AI finally feels like a real app — fast, smooth, usable. ⚡ Turn it on: CLAUDE_CODE_NO_FLICKER=1 claude Small launch. Huge UX upgrade. Terminal-native AI is getting scary good.
0:15

Holy shit...Claude Code just killed one of the most annoying terminal problems — flickering. 🤯 Anthropic quietly shipped NO_FLICKER mode and it changes how AI feels inside the terminal: • No more screen flashing • Smooth scrolling + stable viewport • Mouse support (yes… in terminal) • Constant memory/CPU as chats grow • Cleaner selections (no UI junk copied) This isn’t just a UI tweak. They basically virtualized the entire terminal viewport to control rendering themselves. Result? Terminal AI finally feels like a real app — fast, smooth, usable. ⚡ Turn it on: CLAUDE_CODE_NO_FLICKER=1 claude Small launch. Huge UX upgrade. Terminal-native AI is getting scary good.

Suryansh Tiwari

32,145 Aufrufe • vor 7 Monaten

Grok 4.7 is lazy. I ran it in Cursor for 10 hours, and I gave it the image from every angle. Sad to say the hype is not worth it. Not good enough just yet. The launch benchmarks say the same thing, 26% on Terminal-Bench 4.0. Have to see what Grok 5 looks like.
0:13

Grok 4.7 is lazy. I ran it in Cursor for 10 hours, and I gave it the image from every angle. Sad to say the hype is not worth it. Not good enough just yet. The launch benchmarks say the same thing, 26% on Terminal-Bench 4.0. Have to see what Grok 5 looks like.

Ziwen

10,420 Aufrufe • vor 1 Monat

Been working on CAT for a while. It’s finally live. Cat OS brings Solana to the terminal. Wallets, positions, approvals and scheduled tasks all follow the same command grammar. Everything running locally is free. Persistent execution uses $C4T. CA: 2sLKA6PzFVjeTUaiKBCkDbDiS8MuoZAeXxZMgs9Apump Close the terminal and scheduled tasks keep running onchain through a network of collateralized operators. Full breakdown below.
0:18

Been working on CAT for a while. It’s finally live. Cat OS brings Solana to the terminal. Wallets, positions, approvals and scheduled tasks all follow the same command grammar. Everything running locally is free. Persistent execution uses $C4T. CA: 2sLKA6PzFVjeTUaiKBCkDbDiS8MuoZAeXxZMgs9Apump Close the terminal and scheduled tasks keep running onchain through a network of collateralized operators. Full breakdown below.

anonDev_

111,290 Aufrufe • vor 2 Monaten

martin necas getting pushed into the islanders bench and then randomly punching brock nelson… #cinema
0:11

martin necas getting pushed into the islanders bench and then randomly punching brock nelson… #cinema

allie SAW WHALIEN52&HOME&LOVEMAZE&MAGICSHOP

892,869 Aufrufe • vor 8 Monaten

My Task Agent got an update 🧠 Schedule tasks for your agent to complete every hour or every day
0:19

My Task Agent got an update 🧠 Schedule tasks for your agent to complete every hour or every day

griffain

77,741 Aufrufe • vor 1 Jahr

Introducing Mission Task: The easiest way to earn 1Launch Point! Mission Task is coming soon to 1Launch, and it's about to become your secret weapon for stacking those precious points! What is Mission Task? It's your one-stop shop for earning 1Launch Points through simple social media tasks. Say goodbye to complex hoops and hello to effortless point accumulation! Coming very soon, Mission Tasks will be your one-stop shop for skyrocketing your 1Launch Point balance through engaging social media tasks. In the meantime, drop a comment below! What social media tasks would you love to see in Mission Task? #1Launch #earnings #Airdrop #ComingSoon
0:14

Introducing Mission Task: The easiest way to earn 1Launch Point! Mission Task is coming soon to 1Launch, and it's about to become your secret weapon for stacking those precious points! What is Mission Task? It's your one-stop shop for earning 1Launch Points through simple social media tasks. Say goodbye to complex hoops and hello to effortless point accumulation! Coming very soon, Mission Tasks will be your one-stop shop for skyrocketing your 1Launch Point balance through engaging social media tasks. In the meantime, drop a comment below! What social media tasks would you love to see in Mission Task? #1Launch #earnings #Airdrop #ComingSoon

0xTotally

15,732 Aufrufe • vor 2 Jahren

Claude Code can now send push notifications to your phone when a long task finishes or Claude needs your input. Walk away from the terminal, we'll let you know when it's done.
0:14

Claude Code can now send push notifications to your phone when a long task finishes or Claude needs your input. Walk away from the terminal, we'll let you know when it's done.

ClaudeDevs

1,340,300 Aufrufe • vor 6 Monaten

GPT-5.1 is now live in Warp. It’s much faster (40% faster task completion on a subset of SWE-bench Verified) without compromising quality. GPT-5.1 is available to all Warp users, and is now the default model for all new users.
0:13

GPT-5.1 is now live in Warp. It’s much faster (40% faster task completion on a subset of SWE-bench Verified) without compromising quality. GPT-5.1 is available to all Warp users, and is now the default model for all new users.

Warp

80,683 Aufrufe • vor 1 Jahr

Muscle memory: When a movement is repeated over time, the brain creates a long-term muscle memory for that task, eventually allowing it to be performed with little to no conscious effort.
0:15

Muscle memory: When a movement is repeated over time, the brain creates a long-term muscle memory for that task, eventually allowing it to be performed with little to no conscious effort.

HOW THINGS WORK

7,799,774 Aufrufe • vor 2 Jahren

We’re excited to introduce Text-to-LoRA: a Hypernetwork that generates task-specific LLM adapters (LoRAs) based on a text description of the task. Catch our presentation at #ICML2025! Paper: Code: Biological systems are capable of rapid adaptation, given limited sensory cues. For example, our human visual system can quickly adapt and tune its light sensitivity to our surroundings. While modern LLMs exhibit a wide variety of capabilities and knowledge, they remain rigid when adding task-specific capabilities. Traditionally, customizing these models requires gathering large datasets and performing often expensive, time-consuming fine-tuning for specific applications. To bypass these limitations, Text-to-LoRA (T2L) meta-learns a “hypernetwork” that takes in a text description of a desired task, as a prompt, and generates a task-specific LoRA that performs well on the task. In our experiments, we show that T2L can encode hundreds of existing LoRA adapters. While the compression is lossy, T2L maintains the performance of task-specifically tuned LoRA adapters. We also show that T2L can even generalize to unseen tasks given a natural language description of the tasks. Importantly, Text-to-LoRA is parameter-efficient. It generates LoRAs in a single, inexpensive step, based solely on a simple text description of the task. This approach is a step towards dramatically lowering the technical and computational barriers, allowing non-technical users to specialize foundation models using plain language, rather than needing deep technical expertise or large compute resources.
0:18

We’re excited to introduce Text-to-LoRA: a Hypernetwork that generates task-specific LLM adapters (LoRAs) based on a text description of the task. Catch our presentation at #ICML2025! Paper: Code: Biological systems are capable of rapid adaptation, given limited sensory cues. For example, our human visual system can quickly adapt and tune its light sensitivity to our surroundings. While modern LLMs exhibit a wide variety of capabilities and knowledge, they remain rigid when adding task-specific capabilities. Traditionally, customizing these models requires gathering large datasets and performing often expensive, time-consuming fine-tuning for specific applications. To bypass these limitations, Text-to-LoRA (T2L) meta-learns a “hypernetwork” that takes in a text description of a desired task, as a prompt, and generates a task-specific LoRA that performs well on the task. In our experiments, we show that T2L can encode hundreds of existing LoRA adapters. While the compression is lossy, T2L maintains the performance of task-specifically tuned LoRA adapters. We also show that T2L can even generalize to unseen tasks given a natural language description of the tasks. Importantly, Text-to-LoRA is parameter-efficient. It generates LoRAs in a single, inexpensive step, based solely on a simple text description of the task. This approach is a step towards dramatically lowering the technical and computational barriers, allowing non-technical users to specialize foundation models using plain language, rather than needing deep technical expertise or large compute resources.

Sakana AI

403,451 Aufrufe • vor 1 Jahr

MiniMax M2.7 is live on Runware on Day0! 🚀 - full end-to-end project delivery - self-evolving architecture: 30% performance gains across 100+ iterations - 97% skill adherence rate across 40+ complex tasks - 56.22% on SWE-Pro, 55.6% on VIBE-Pro, 57.0% on Terminal Bench 2
0:18

MiniMax M2.7 is live on Runware on Day0! 🚀 - full end-to-end project delivery - self-evolving architecture: 30% performance gains across 100+ iterations - 97% skill adherence rate across 40+ complex tasks - 56.22% on SWE-Pro, 55.6% on VIBE-Pro, 57.0% on Terminal Bench 2

Runware

38,628 Aufrufe • vor 7 Monaten

Ref loses a blade and hitches a lift to the bench 🤣
0:13

Ref loses a blade and hitches a lift to the bench 🤣

Behind The Net

10,929 Aufrufe • vor 8 Monaten

That’s a WRAP, on The Terminal List Season 2 production. From South Africa, to North America, to Morocco, the cast and crew pushed the limits to bring book 2, TRUE BELIEVER to life!
0:14

That’s a WRAP, on The Terminal List Season 2 production. From South Africa, to North America, to Morocco, the cast and crew pushed the limits to bring book 2, TRUE BELIEVER to life!

The Terminal List

191,161 Aufrufe • vor 1 Jahr