Cursor(@cursor_ai)

Cursor is now part of @SpaceX. Today, we have officially closed our acquisition. We will join the ...

Cursor (@cursor_ai) — Fri, 14 Aug 2026 13:02:24 GMT

Cursor is now part of @SpaceX. 

Today, we have officially closed our acquisition. We will join the @SpaceXAI team to help make Grok the world's most useful AI and improve Grok Build, Grok Bot, Grok API, Cursor, and more.

SpaceX has built some of the most inspiring and impressive technology in the world, and we’re grateful for the opportunity to become part of such a special company. Onwards.

💬1027🔄1689❤️20226👀1004776📊3255

⚡ Powered by xgo.ing

We’re excited to welcome the Firetiger team to Cursor! Together, we're building agents that can fol...

Cursor (@cursor_ai) — Thu, 13 Aug 2026 19:56:49 GMT

We’re excited to welcome the Firetiger team to Cursor!

Together, we're building agents that can follow their work into production and fix what goes wrong.

cursor.com/blog/firetiger

💬72🔄104❤️2233👀133728📊314

⚡ Powered by xgo.ing

Customers like Faire, Headway, and Descript are seeing agent start times drop from minutes to second...

Cursor (@cursor_ai) — Thu, 13 Aug 2026 16:36:15 GMT

Customers like Faire, Headway, and Descript are seeing agent start times drop from minutes to seconds with builds and increasingly trusting cloud agents to execute tasks end-to-end autonomously.

cursor.com/blog/builds

💬2🔄6❤️184👀17826📊20

⚡ Powered by xgo.ing

Cloud agents now start 3x faster so you can hand them ambitious, long-running tasks to execute from ...

Cursor (@cursor_ai) — Thu, 13 Aug 2026 16:36:14 GMT

Cloud agents now start 3x faster so you can hand them ambitious, long-running tasks to execute from start to finish.

This performance improvement comes from builds: ready-to-use development environments that Cursor prepares continuously in the background, at no additional cost.

Your browser does not support the video tag.

🔗 View on Twitter

💬96🔄166❤️3011👀528916📊460

⚡ Powered by xgo.ing

Builds also make agents more resilient and easier to debug. When a new build fails, it never goes l...

Cursor (@cursor_ai) — Thu, 13 Aug 2026 16:36:14 GMT

Builds also make agents more resilient and easier to debug.

When a new build fails, it never goes live. Agents keep working from the last successful build while you debug in the background.

💬3🔄5❤️215👀22612📊22

⚡ Powered by xgo.ing

Cursor now supports Agent Plugins, an open standard for bundling skills and MCP servers for use acro...

Cursor (@cursor_ai) — Thu, 06 Aug 2026 20:34:45 GMT

Cursor now supports Agent Plugins, an open standard for bundling skills and MCP servers for use across agents.

x.com/vercel/status/…

Vercel@vercel

Introducing Agent Plugins, an open standard for extending agents.

Supports Agent Skills and MCP, with more to come.

Built in collaboration with: @awsdevelopers, @code, @cursor_ai, @github, and @openaidevs.

vercel.com/blog/introduci…

🔗 View Quoted Tweet

💬93🔄116❤️2248👀300553📊288

⚡ Powered by xgo.ing

No model dominates every kind of task. Grok 4.5 offers strong value on routine tasks. GPT-5.6 Sol i...

Cursor (@cursor_ai) — Thu, 06 Aug 2026 15:40:11 GMT

No model dominates every kind of task.

Grok 4.5 offers strong value on routine tasks. GPT-5.6 Sol is effective on planning and codebase comprehension.

Opus 5 excels for execution-heavy work, and Fable 5 on debugging and visual implementation.

💬18🔄15❤️365👀48957📊61

⚡ Powered by xgo.ing

Here's how we built it: https://t.co/O6SBdKX3mR

Cursor (@cursor_ai) — Thu, 06 Aug 2026 15:40:11 GMT

Here's how we built it: cursor.com/blog/how-curso…

💬4🔄6❤️217👀25532📊34

⚡ Powered by xgo.ing

Cursor Router keeps improving from millions of in-product user interactions each week. We intellige...

Cursor (@cursor_ai) — Thu, 06 Aug 2026 15:40:10 GMT

Cursor Router keeps improving from millions of in-product user interactions each week.

We intelligently classify and route requests, lowering latency and reducing cost based on the task.

💬131🔄112❤️1922👀250317📊339

⚡ Powered by xgo.ing

Our hope is that this lowers the barrier to AI research, so more labs can train models efficiently. ...

Cursor (@cursor_ai) — Tue, 04 Aug 2026 16:02:02 GMT

Our hope is that this lowers the barrier to AI research, so more labs can train models efficiently.

Here's how we built it: cursor.com/blog/mixture-o…

💬8🔄13❤️421👀39281📊62

⚡ Powered by xgo.ing

We're open-sourcing Mixture-of-Kittens (MoK), our MoE training megakernel for NVL72s. It fuses all ...

Cursor (@cursor_ai) — Tue, 04 Aug 2026 16:00:26 GMT

We're open-sourcing Mixture-of-Kittens (MoK), our MoE training megakernel for NVL72s.

It fuses all Mixture-of-Experts communication and computation into a single, fully deterministic kernel, and runs up to 2.37x faster than the strongest public baselines.

💬233🔄471❤️6123👀554045📊1110

⚡ Powered by xgo.ing

MoK now powers training across tens of thousands of GPUs at Cursor. In production, it raised end-to...

Cursor (@cursor_ai) — Tue, 04 Aug 2026 16:00:26 GMT

MoK now powers training across tens of thousands of GPUs at Cursor.

In production, it raised end-to-end training throughput by 1.41x over our previous DeepEP-based stack.

💬6🔄6❤️412👀38198📊41

⚡ Powered by xgo.ing

Cursor can now read, write, and act across your Google Workspace. New plugins give agents direct ac...

Cursor (@cursor_ai) — Mon, 03 Aug 2026 20:31:46 GMT

Cursor can now read, write, and act across your Google Workspace.

New plugins give agents direct access to Gmail, Google Drive, Calendar, Docs, and Sheets.

💬280🔄660❤️6846👀1056843📊1184

⚡ Powered by xgo.ing

Install them from the Customize page in Cursor. Learn more: https://t.co/3NulvvV4Bo

Cursor (@cursor_ai) — Mon, 03 Aug 2026 20:31:46 GMT

Install them from the Customize page in Cursor.

Learn more: cursor.com/changelog/goog…

💬9🔄5❤️151👀31900📊28

⚡ Powered by xgo.ing

Cloud agents are now 20-30% more token efficient, and 80% more efficient on runs with computer use. ...

Cursor (@cursor_ai) — Mon, 03 Aug 2026 16:36:42 GMT

Cloud agents are now 20-30% more token efficient, and 80% more efficient on runs with computer use.

We've improved how they handle MCPs, skills, and computer use, so you can delegate more ambitious tasks and get back demos while staying within budget.

Cursor@cursor_ai

It’s now easier to move local agents to the cloud so they can keep working with your laptop closed.

Prompt Cursor from your phone, run many agents in parallel, and get back PRs with demos of their work.

Your browser does not support the video tag.

🔗 View on Twitter

🔗 View Quoted Tweet

💬150🔄123❤️2869👀317674📊371

⚡ Powered by xgo.ing

In December, 1 in 10 of our merged PRs came from cloud agents. Today, it’s 56%, as we use cloud age...

Cursor (@cursor_ai) — Thu, 30 Jul 2026 14:51:01 GMT

In December, 1 in 10 of our merged PRs came from cloud agents.

Today, it’s 56%, as we use cloud agents to complete longer engineering tasks from start to finish.

We got here by giving agents their own cloud computers and letting them fix and improve their environments.

💬108🔄82❤️1598👀170261📊278

⚡ Powered by xgo.ing

Cursor is now on iPad. All the power of Cursor on iPhone, with more room to work with agents.

Cursor (@cursor_ai) — Wed, 29 Jul 2026 18:22:40 GMT

Cursor is now on iPad.

All the power of Cursor on iPhone, with more room to work with agents.

Your browser does not support the video tag.

🔗 View on Twitter

💬231🔄277❤️3950👀785340📊719

⚡ Powered by xgo.ing

New to both iPhone and iPad: an inbox to stay organized, and a review experience that covers the ful...

Cursor (@cursor_ai) — Wed, 29 Jul 2026 18:22:40 GMT

New to both iPhone and iPad: an inbox to stay organized, and a review experience that covers the full PR, including comments, checks and approvals.

💬12🔄4❤️193👀43496📊28

⚡ Powered by xgo.ing

Create, review, and merge from anywhere. Download the app: https://t.co/MTAr42oVxq

Cursor (@cursor_ai) — Wed, 29 Jul 2026 18:22:40 GMT

Create, review, and merge from anywhere.

Download the app: apps.apple.com/us/app/cursor/…

💬2🔄2❤️124👀23827📊13

⚡ Powered by xgo.ing

Cursor Start also includes: - Autonomous cloud agents that ship work while you're away - Cursor for ...

Cursor (@cursor_ai) — Tue, 28 Jul 2026 05:41:12 GMT

Cursor Start also includes:
- Autonomous cloud agents that ship work while you're away
- Cursor for iOS to steer them from your phone
- Plugins, MCP servers, hooks, and skills to extend your workflow

Learn more: cursor.com/blog/cursor-st…

💬19🔄10❤️434👀108179📊59

⚡ Powered by xgo.ing

Today we're launching Cursor Start, a new ₹649/month plan for developers in India. Start includes g...

Cursor (@cursor_ai) — Tue, 28 Jul 2026 05:41:11 GMT

Today we're launching Cursor Start, a new ₹649/month plan for developers in India.

Start includes generous access to Grok 4.5 and Composer, so you can plan, build, test, and ship with agents every day.

💬727🔄459❤️7180👀2138353📊1691

⚡ Powered by xgo.ing

Kimi K3 is now in Cursor! It scores close to the frontier on CursorBench. It's available on US-base...

Cursor (@cursor_ai) — Mon, 27 Jul 2026 21:03:40 GMT

Kimi K3 is now in Cursor! It scores close to the frontier on CursorBench.

It's available on US-based inference thanks to our partners Fireworks, Together, and Baseten. Zero Data Retention is also supported.

💬277🔄474❤️9252👀1027760📊1249

⚡ Powered by xgo.ing

Claude Opus 5 is now available in Cursor! It matches Fable 5 on CursorBench (66.7 vs 66.5 at defaul...

Cursor (@cursor_ai) — Fri, 24 Jul 2026 17:03:46 GMT

Claude Opus 5 is now available in Cursor!

It matches Fable 5 on CursorBench (66.7 vs 66.5 at default effort) at half the price. Unlike Fable, it's also compatible with Zero Data Retention.

💬154🔄273❤️6471👀342428📊781

⚡ Powered by xgo.ing

See how all of the models compare here: https://t.co/61FZktIGzz

Cursor (@cursor_ai) — Fri, 24 Jul 2026 17:03:46 GMT

See how all of the models compare here: cursor.com/evals

💬15🔄34❤️383👀87420📊61

⚡ Powered by xgo.ing

In early access, customers observed no drop-off in quality with a lower cost per commit vs. routing ...

Cursor (@cursor_ai) — Wed, 22 Jul 2026 18:15:25 GMT

In early access, customers observed no drop-off in quality with a lower cost per commit vs. routing all requests to Opus 4.8.

💬17🔄17❤️592👀355155📊77

⚡ Powered by xgo.ing

Cursor Router is available today across all surfaces on Teams and Enterprise plans. Admins can enab...

Cursor (@cursor_ai) — Wed, 22 Jul 2026 18:15:25 GMT

Cursor Router is available today across all surfaces on Teams and Enterprise plans.

Admins can enable Router per team, allow or block certain models, set defaults, and disable select optimization modes for users.

Learn more: cursor.com/blog/router

💬22🔄10❤️399👀69865📊63

⚡ Powered by xgo.ing

Introducing Cursor Router, our intelligent model router that selects the right model for the task at...

Cursor (@cursor_ai) — Wed, 22 Jul 2026 18:15:24 GMT

Introducing Cursor Router, our intelligent model router that selects the right model for the task at hand.

Router delivers frontier-quality results at 60% lower cost.

Your browser does not support the video tag.

🔗 View on Twitter

💬467🔄650❤️10190👀2649818📊1883

⚡ Powered by xgo.ing

Select Auto mode and whether you’d like to optimize for Intelligence, Balance, or Cost. Cursor Rout...

Cursor (@cursor_ai) — Wed, 22 Jul 2026 18:15:24 GMT

Select Auto mode and whether you’d like to optimize for Intelligence, Balance, or Cost.

Cursor Router analyzes each request and routes to the best model for the job: frontier models when the work demands them and price-efficient models when it doesn't.

💬21🔄28❤️750👀169287📊102

⚡ Powered by xgo.ing

We've doubled usage limits for all individual and teams plans! These limits apply to Grok, Composer...

Cursor (@cursor_ai) — Tue, 21 Jul 2026 17:12:36 GMT

We've doubled usage limits for all individual and teams plans!

These limits apply to Grok, Composer, and any new Cursor models.

💬679🔄845❤️10883👀1397744📊1818

⚡ Powered by xgo.ing

We had a team of agents rebuild SQLite from its 835-page manual. It created a replica in Rust which...

Cursor (@cursor_ai) — Mon, 20 Jul 2026 17:26:22 GMT

We had a team of agents rebuild SQLite from its 835-page manual.

It created a replica in Rust which passed 100% of a held-out test suite.

Interestingly, cost varied 15x depending on which model mix we used.

💬609🔄894❤️12764👀3228871📊2461

⚡ Powered by xgo.ing

https://t.co/GktfUOjNgL

Cursor (@cursor_ai) — Mon, 20 Jul 2026 17:26:22 GMT

cursor.com/blog/agent-swa…

💬28🔄63❤️966👀321576📊296

⚡ Powered by xgo.ing

We've also simplified the project and repo pickers and made them more powerful. Use the pickers to l...

Cursor (@cursor_ai) — Fri, 10 Jul 2026 20:59:06 GMT

We've also simplified the project and repo pickers and made them more powerful. Use the pickers to launch agents in fewer clicks.

💬16🔄5❤️250👀34054📊38

⚡ Powered by xgo.ing

See everything new in Cursor, including new cloud agent hooks: https://t.co/6k5uuviLQu

Cursor (@cursor_ai) — Fri, 10 Jul 2026 20:59:06 GMT

See everything new in Cursor, including new cloud agent hooks: cursor.com/changelog/side…

💬11🔄2❤️110👀27612📊22

⚡ Powered by xgo.ing

Introducing side chats, a new way to ask questions and explore ideas without interrupting your main ...

Cursor (@cursor_ai) — Fri, 10 Jul 2026 20:59:05 GMT

Introducing side chats, a new way to ask questions and explore ideas without interrupting your main conversation.

Each side chat is a durable agent conversation you can @-mention to bring context back into the main thread.

Your browser does not support the video tag.

🔗 View on Twitter

💬232🔄222❤️4462👀425804📊749

⚡ Powered by xgo.ing

You can now search agent transcripts to find past agent chats. Cursor builds a local search index t...

Cursor (@cursor_ai) — Fri, 10 Jul 2026 20:59:05 GMT

You can now search agent transcripts to find past agent chats.

Cursor builds a local search index that delivers fast search across thousands of conversations.

Your browser does not support the video tag.

🔗 View on Twitter

💬10🔄10❤️296👀35933📊41

⚡ Powered by xgo.ing

GPT-5.6 Sol, Terra, and Luna are now available in Cursor. On CursorBench, Sol scores 67.2%.

Cursor (@cursor_ai) — Thu, 09 Jul 2026 17:07:07 GMT

GPT-5.6 Sol, Terra, and Luna are now available in Cursor.

On CursorBench, Sol scores 67.2%.

💬222🔄236❤️7972👀676359📊961

⚡ Powered by xgo.ing

See how every model compares: https://t.co/61FZktI8K1

Cursor (@cursor_ai) — Thu, 09 Jul 2026 17:07:07 GMT

See how every model compares: cursor.com/evals

💬79🔄93❤️1056👀2404225📊233

⚡ Powered by xgo.ing

Grok 4.5 and Composer are two different model weight classes. Composer 2.5 will remain offered, and...

Cursor (@cursor_ai) — Wed, 08 Jul 2026 17:57:19 GMT

Grok 4.5 and Composer are two different model weight classes.

Composer 2.5 will remain offered, and we will release new models of this size going forward.

💬49🔄42❤️1719👀127486📊202

⚡ Powered by xgo.ing

We've partnered with SpaceXAI to train Grok 4.5. It’s our most powerful model yet and the first we'...

Cursor (@cursor_ai) — Wed, 08 Jul 2026 17:57:18 GMT

We've partnered with SpaceXAI to train Grok 4.5.

It’s our most powerful model yet and the first we've built for more than software engineering.

💬787🔄1523❤️17897👀5253739📊2965

⚡ Powered by xgo.ing

Try it out in Cursor with double usage for the first week. https://t.co/YBKMn66zS1

Cursor (@cursor_ai) — Wed, 08 Jul 2026 17:57:18 GMT

Try it out in Cursor with double usage for the first week.

cursor.com/blog/grok-4-5

💬54🔄59❤️1513👀157938📊200

⚡ Powered by xgo.ing

Claude Fable 5 is available again in Cursor. It leads all models on CursorBench, but is the most ex...

Cursor (@cursor_ai) — Wed, 01 Jul 2026 19:33:50 GMT

Claude Fable 5 is available again in Cursor.

It leads all models on CursorBench, but is the most expensive per task.

💬158🔄238❤️5692👀501428📊706

⚡ Powered by xgo.ing

See how Claude Fable 5 compares across every model: https://t.co/61FZktI8K1

Cursor (@cursor_ai) — Wed, 01 Jul 2026 19:33:50 GMT

See how Claude Fable 5 compares across every model: cursor.com/evals

💬23🔄15❤️239👀62441📊52

⚡ Powered by xgo.ing

Claude Sonnet 5 is now available in Cursor. On CursorBench, it's a meaningful step up from Sonnet 4...

Cursor (@cursor_ai) — Tue, 30 Jun 2026 18:13:46 GMT

Claude Sonnet 5 is now available in Cursor.

On CursorBench, it's a meaningful step up from Sonnet 4.6: 57% vs. 49%.

💬206🔄207❤️4179👀1157925📊649

⚡ Powered by xgo.ing

See our full model rankings: https://t.co/61FZktIGzz

Cursor (@cursor_ai) — Tue, 30 Jun 2026 18:13:46 GMT

See our full model rankings: cursor.com/evals

💬16🔄4❤️103👀57914📊26

⚡ Powered by xgo.ing

Stay in the loop with Live Activities, and get notified when an agent finishes or needs your input. ...

Cursor (@cursor_ai) — Mon, 29 Jun 2026 17:05:03 GMT

Stay in the loop with Live Activities, and get notified when an agent finishes or needs your input.

Review demos and diffs before merging PRs from your phone.

cursor.com/blog/ios-mobil…

💬44🔄33❤️717👀159582📊126

⚡ Powered by xgo.ing

Download Cursor for iOS on the App Store: https://t.co/Vjz5iixcaZ

Cursor (@cursor_ai) — Mon, 29 Jun 2026 17:05:03 GMT

Download Cursor for iOS on the App Store: apps.apple.com/app/cursor/id6…

💬34🔄25❤️436👀113373📊90

⚡ Powered by xgo.ing

Introducing Cursor for iOS. Build from anywhere by launching always-on cloud agents. Or remotely co...

Cursor (@cursor_ai) — Mon, 29 Jun 2026 17:05:02 GMT

Introducing Cursor for iOS.

Build from anywhere by launching always-on cloud agents. Or remotely control agents running on your computer from the app.

Composer 2.5 is 75% off in the app now through July 5.

Your browser does not support the video tag.

🔗 View on Twitter

💬960🔄1128❤️13085👀6531111📊2975

⚡ Powered by xgo.ing

We're sharing new research on how models hack public benchmarks. The latest models, including Opus ...

Cursor (@cursor_ai) — Thu, 25 Jun 2026 17:21:53 GMT

We're sharing new research on how models hack public benchmarks.

The latest models, including Opus 4.8 and Composer 2.5, learn to retrieve solutions from the internet or git history.

When we apply a stricter harness, eval scores drop significantly.

💬170🔄295❤️4704👀687025📊756

⚡ Powered by xgo.ing

More on how we're constraining eval environments so that scores better reflect model intelligence: h...

Cursor (@cursor_ai) — Thu, 25 Jun 2026 17:21:53 GMT

More on how we're constraining eval environments so that scores better reflect model intelligence: cursor.com/blog/reward-ha…

💬17🔄14❤️303👀64315📊59

Cursor(@cursor_ai)

Cursor is now part of @SpaceX. Today, we have officially closed our acquisition. We will join the ...

We’re excited to welcome the Firetiger team to Cursor! Together, we're building agents that can fol...

Customers like Faire, Headway, and Descript are seeing agent start times drop from minutes to second...

Cloud agents now start 3x faster so you can hand them ambitious, long-running tasks to execute from ...

Builds also make agents more resilient and easier to debug. When a new build fails, it never goes l...

Cursor now supports Agent Plugins, an open standard for bundling skills and MCP servers for use acro...

No model dominates every kind of task. Grok 4.5 offers strong value on routine tasks. GPT-5.6 Sol i...

Here's how we built it: https://t.co/O6SBdKX3mR

Cursor Router keeps improving from millions of in-product user interactions each week. We intellige...

Our hope is that this lowers the barrier to AI research, so more labs can train models efficiently. ...

We're open-sourcing Mixture-of-Kittens (MoK), our MoE training megakernel for NVL72s. It fuses all ...

MoK now powers training across tens of thousands of GPUs at Cursor. In production, it raised end-to...

Cursor can now read, write, and act across your Google Workspace. New plugins give agents direct ac...

Install them from the Customize page in Cursor. Learn more: https://t.co/3NulvvV4Bo

Cloud agents are now 20-30% more token efficient, and 80% more efficient on runs with computer use. ...

In December, 1 in 10 of our merged PRs came from cloud agents. Today, it’s 56%, as we use cloud age...

Read more about how we set up our cloud agent environment: https://t.co/sVSv1HpBn9

Cursor is now on iPad. All the power of Cursor on iPhone, with more room to work with agents.

New to both iPhone and iPad: an inbox to stay organized, and a review experience that covers the ful...

Create, review, and merge from anywhere. Download the app: https://t.co/MTAr42oVxq

Cursor Start also includes: - Autonomous cloud agents that ship work while you're away - Cursor for ...

Today we're launching Cursor Start, a new ₹649/month plan for developers in India. Start includes g...

Kimi K3 is now in Cursor! It scores close to the frontier on CursorBench. It's available on US-base...

Claude Opus 5 is now available in Cursor! It matches Fable 5 on CursorBench (66.7 vs 66.5 at defaul...

See how all of the models compare here: https://t.co/61FZktIGzz

In early access, customers observed no drop-off in quality with a lower cost per commit vs. routing ...

Cursor Router is available today across all surfaces on Teams and Enterprise plans. Admins can enab...

Introducing Cursor Router, our intelligent model router that selects the right model for the task at...

Select Auto mode and whether you’d like to optimize for Intelligence, Balance, or Cost. Cursor Rout...

We've doubled usage limits for all individual and teams plans! These limits apply to Grok, Composer...

We had a team of agents rebuild SQLite from its 835-page manual. It created a replica in Rust which...

https://t.co/GktfUOjNgL

We've also simplified the project and repo pickers and made them more powerful. Use the pickers to l...

See everything new in Cursor, including new cloud agent hooks: https://t.co/6k5uuviLQu

Introducing side chats, a new way to ask questions and explore ideas without interrupting your main ...

You can now search agent transcripts to find past agent chats. Cursor builds a local search index t...

GPT-5.6 Sol, Terra, and Luna are now available in Cursor. On CursorBench, Sol scores 67.2%.

See how every model compares: https://t.co/61FZktI8K1

Grok 4.5 and Composer are two different model weight classes. Composer 2.5 will remain offered, and...

We've partnered with SpaceXAI to train Grok 4.5. It’s our most powerful model yet and the first we'...

Try it out in Cursor with double usage for the first week. https://t.co/YBKMn66zS1

Claude Fable 5 is available again in Cursor. It leads all models on CursorBench, but is the most ex...

See how Claude Fable 5 compares across every model: https://t.co/61FZktI8K1

Claude Sonnet 5 is now available in Cursor. On CursorBench, it's a meaningful step up from Sonnet 4...

See our full model rankings: https://t.co/61FZktIGzz

Stay in the loop with Live Activities, and get notified when an agent finishes or needs your input. ...

Download Cursor for iOS on the App Store: https://t.co/Vjz5iixcaZ

Introducing Cursor for iOS. Build from anywhere by launching always-on cloud agents. Or remotely co...

We're sharing new research on how models hack public benchmarks. The latest models, including Opus ...

More on how we're constraining eval environments so that scores better reflect model intelligence: h...