Anton

Today’s RepoSuite progress: 1) api service containers split brainstorm. The idea is to create 4 services: api,db,wss & nats for api <-> wss communications. db: PostgreSQL, api/wss/nats: Go 2) implemented json-schema. org for shared json interfaces for usage in api/ui/wss (this will be single repo with the single source of truth purpose). I've had to rename my db repo to schema repo - now it is single repo that holds db & json schema; db schema uses Liquibase btw. 3) db role now renamed to schema_master, after db repo refactor 4) initial db refactor started to support workbench -> workspace -> panel flow; terminal panel json schema created for state column; not quite sure if it is final - because now it uses mixed fields with agent session... probably there should be panel that is extended / uses base type terminal for agents... not quite sure yet... future tasks: identify what is agent and how it should be stored in db (single agent table or agent types + agent sessions), and how terminal with agent should be registered correctly (probably as subtype/base from terminal - because it will require custom startup logic/scripting - like selected harness (codex/opencode/etc) and session + custom ui extras compared to terminal...

Qwen3.6 35B-A3B dropped yesterday, so I ran it on 4 GPUs to see how it performs: 🟣 RTX 3090 — 49.78 tok/s, TTFT 852ms 🟡 RTX 4090 — 118.93 tok/s, TTFT 686ms 🟢 RTX 5090 — 160.37 tok/s, TTFT 409ms 🔵 DGX Spark — 59.98 tok/s, TTFT 228ms I went with ollama as the backend because honestly, it's the easiest way for most people to get started. One command, model pulled, done. I used Q4_K_M (24GB) across all four cards. The reason is the 3090 and 4090 don't support NVFP4 (only the 5090 and DGX Spark could use it). Keeping the same quant everywhere felt like the fairest way to compare. And yes, you can absolutely squeeze more performance out of every card with vLLM, SGLang, or TensorRT-LLM. But that's not what this test is about. This is just the out-of-the-box experience for folks who own a GPU and want to try the new model tonight.

Congrats! Opus 4.7 is less than 4% better than GPT-5.4 (already a month old) while costing 2x more Such a huge update!

@alexwtlf Switched from Who?

Has anyone actually switched from Claude Code to Codex?

Today's challenge 💪

nerfing ai models and then rereleasing them with a new version number

Shocking result on my pelican benchmark this morning, I got a better pelican from a 21GB local Qwen3.6-35B-A3B running on my laptop than I did from the new Opus 4.7! Qwen on the left, Opus on the right

Serious question. Has anyone ever noticed meaningful regressions in Codex/OpenAI models? I feel like we talk about this a lot w/ Anthropic but I've never seen a similar discussion with OAI.

> Opus 4.7 is a ~5T model > Qwen 3.6 uses 3B for inference SWE Bench verified: > Opus 4.7: 87.6% > Qwen3.6-35B-A3B: 73.4% No rate limits. Free to run. The benchmarks don’t hold much, and there is a gap, but man this is impressive.

Qwen3.6-35B-A3B-TQ3_4S Smarter 💡 Faster (PP512 = 1832 tok/s, TG128 = 107 tok/s) 🚀 Smaller 12.4 GiB 🤏 huggingface.co/YTan2000/Qwen3…

I am happy everyone is switching to Codex, but Tibo if you start rate limiting me or making me use worse models...

Hi! To celebrate its 1-year anniversary, I have allowed Codex to reset its own rate limits across all plans. Enjoy all the new features.

Be Anthropic > Give people Opus 4.6 > People love it. > For 2 months you degrade Opus 4.6 > You give back normal Opus 4.6 and call it Opus 4.7. > People love it. That's the business model.

2-bit Qwen3.6-35B-A3B did a complete repo bug hunt with evidence, repro, fixes, tests and a PR writeup. 🔥 Run it locally in Unsloth Studio with just 13GB RAM. 2-bit Qwen3.6 GGUF made 30+ tool calls, searched 20 sites and executed Python code. GitHub: github.com/unslothai/unsl…

Three.js r184 released ⛅️ threejs.org/changelog/?r184

Codex just got a lot more powerful. Computer use, in-app browser, image generation and editing, 90+ new plugins to connect to everything, multi-terminal, SSH into devboxes, thread automations, rich document editing. Learns from experience and proactively suggestions work. And a ton more.

The wait is over. Cloudflare Email Service is now in public beta 📧 Send and receive emails directly from Workers or REST API with global delivery on Cloudflare's network And just in time for you to build email agents with the Agents SDK!

I cannot wait for gpt 5.5 here in ~30 minutes

Qwen3.6-35B-A3B can now be run locally!💜 The model is the strongest mid-sized LLM on nearly all benchmarks. Run on 23GB RAM via Unsloth Dynamic GGUFs. GGUFs to run: huggingface.co/unsloth/Qwen3.… Guide: unsloth.ai/docs/models/qw…

@thsottiaux @dkundel New codex and reset ?

Feeling codexy today

@Alibaba_Qwen Thank you for supporting open-source! We just made GGUFs so you can run the model locally on 23GB RAM / Mac for 4-bit. 💜🥰 Qwen3.6-35-A3B: huggingface.co/unsloth/Qwen3.…