Amin I.

Find web design leads & businesses without websites effortlessly from any location with just a few clicks. Then contact them with a push of a button: webleadr.com #webdesign #WebDevelopment #leadgeneration

@zivdotcat Cursorbench?? What?

POV: Cursor just dropped its own AI model, Composer 2, already beating Claude Opus 4.6 at coding at half the price.

@PrajwalTomar_ Need SWE bench

Cursor just dropped Composer 2 and it's honestly insane. Their own coding model. 61.7% on Terminal-Bench 2.0, which beats Claude Opus 4.6 sitting at 58.0%. But here's what actually matters: $0.50 per million tokens vs Opus at $5.00. That's 10x cheaper. Sure, GPT-5.4 still dominates at 75.1%, but Composer 2 costs 1/5th the price. For most use cases, that tradeoff is worth it. The secret: They trained it ONLY on code. No random training data. No essays. Just pure code. Plus they built self-summarization into it. Takes 100k token sessions and compresses them down to 1k without losing context. Pretty cool.

@wesbos In that case, using Claude inside Claude Code would give you better results anyway.

Composer 2 vs Opus 4.6 vs GPT 5.4 - a totally unscientific test > Create a Twitter clone. Use Better Auth, Vite, Sqlite, Drizzle, Typescript, and React with Tanstack Start. Each one took ~5 mins in plan mode. Each had access to a browser to test. Composer: 5 mins $6.04 1,250 LOC Opus: 19 mins $10.43 1,000 LOC GPT: 22 mins $14.15 2,000 LOC Opus seems to use the cache WAY more than composer, so it's not really 10x more expensive. Composer app ran first try. Other two needed a bit of CORS debugging but did work. Code between all was extremely similar All three done inside cursor - so Claude Code / Codex may have been different Models used: Composer 2.0 (regular, not fast) Opus 4.6 Medium Thinking GPT 5.4 Medium Thinking

@bridgemindai I only trust Bridgebench.

MiniMax M2.7 just released. #1 on Multi-SWE Bench. 52.7. Above Claude Opus 4.6, Claude Sonnet 4.6, and GPT 5.4. Benchmarks are benchmarks though. Need to see how it holds up in real world vibe coding workflows.

@cryptopunk7213 CursorBench? is this some kind of a buzzword? I am more interested in SWE bench please.

1. theres no way this beats opus 4.6 2. wtf is this chart crime??? 😂

@bridgemindai Based on your post it seems that Composer 2 outperforms Claude Opus in Terminal Bench but what about SWE bench?

Composer 2 outperforms Claude Opus 4.6 Composer 2 scores 61.7 on Terminal-Bench 2.0. Claude Opus 4.6 at 58.0. $0.50/M input. $2.50/M output. 10x cheaper than Claude Opus 4.6. Cursor isn't just an IDE anymore. They're training their own models now. The IDE companies are becoming model companies.

@bridgemindai @claudeai LFG

The new Claude Opus 4.6 1M context window is insane. I am able to launch subagents a lot more now. I used to run into issues with auto compacting which led to hallucinations and lost context. This is a substantial improvement from @claudeai

@uiux_harshit Human

Google Stitch vs Claude vs Human Who is the winner here?

For those using Polar.sh for payments/invoices/taxes, etc. Are you satisfied with it? I am using Paddle and I am interested in using Polar.

This is HUGE! My app for finding web design leads and businesses without websites is almost live! I'm very proud to have implemented the features it needed for my fellow web designers and developers alike just like me. Check out this video! The system detects bad designs, no HTTPS, businesses without websites, pretty much anything to save you (as a fellow developer, SEO specialist, web designer, marketer, etc.) a tremendous amount of time finding ideal web design clients! Let's GO! Stay tuned! #WebDev #buildinpublic #webleadr #webdevelopment #webdesign

@bridgemindai What's also problematic is that I get other obsolete features like extra storage on Google Drive, even though I don't use it. This justifies them making it more expensive. I only want Gemini usage on a subscription basis.

Google won't refund a single dollar on my $250/month AI Ultra subscription. I paid for what I thought was a reliable, premium AI product. Gemini 3.1 Pro ranked #22 on BridgeBench. Gemini CLI crashes mid-session. Antigravity can't hold context. $250/month and they won't even prorate a refund. Just "subscription ends April 12." No refunds on a product that doesn't deliver is insane.

Indeed. Better to trust a legit guy on YouTube testing things out and showing real results right in the video. However, I do use GLM-5 with the Pro Coding Plan subscription from z.ai, inside OpenCode. And I must say the results are actually impressive unlike in your video on z.ai. But the problem might be the free usage that OpenRouter/OpenCode provided for GLM-5 for a limited time?

@webdevamin Appreciate that. Real-world testing over marketing hype every time.

Just cancelled my $250/month Google AI Ultra subscription. Gemini 3.1 Pro is inconsistent. Gemini CLI crashes mid-session. Antigravity loses context on complex tasks. $250/month for a model I can't rely on. Not worth it. Claude Opus 4.6 and GPT 5.4 stay on the stack. Gemini doesn't.

@bridgemindai In BridgeBench we trust

BridgeVoice is here 🔥 Voice-to-text for vibe coding: - Instant on-device Whisper (private, offline) - Cloud mode: 100+ languages - Custom dictionary & instructions for perfect prompts I've used it for 63k+ words. Watch the demo. 50% off first 3 months BridgeMind Pro: BRIDGEVOICEFOUNDER → bridgemind.ai Voice > typing? Reply yes/no 👇

@mscode07 webdevamin.com

Share your website/project 👇

@kaif9999 Just use z.ai/subscribe in OpenCode. Choose the Pro Plan. Lite Plan won't do. This model is on par with the Opus 4.5. This is coming from my own experience.

CLAUDE IS CRASHING MAYBE DOWN AGAIN what do i do now ?

@VadimStrizheus z.ai/subscribe Well, extremely well of course make sure that you’re opt for the problem and not the light plan. It is really working well especially inside OpenCode.

CLAUDE IS CRASHING AND NOT WORKING FOR ME what am I supposed to do?

GLM-5 coding plan subscription and use it inside OpenCode. It is actually on par with Claude Opus 4.5 or even Claude Sonnet 4.6. This comes from my own experience, and I'm really satisfied with the GLM-5 coding plan. Do not use the API key. better opt for the subscription coding plan from ZAI. Choose the Pro plan, not the Light plan, because the Pro plan works way better than the Light plan.

Developers : Is there ANY AI that can actually replace Claude for coding…

How to promote your business for free: 1. Go to HackerNews. 2. Make an account and tap Submit. 3. Write a description of your business. 4. Include your website URL. 5. Submit. Repeat twice a month...

Today we're introducing the world's first AI CMO. Enter your website and it deploys a team of agents to help you get traffic and users. Try it now at okara.ai/cmo