ollama (@ollama) / X

Pinned

GLM-5.2 on Ollama's cloud just got more capacity in US & Europe! Ollama's cloud for GLM 5.2 consistently delivers between 80 to 120 output tokens per second, even during peak hours, compared to 30 to 40 tok/s on other providers. Use GLM-5.2 with Claude Code: ollama launch

Replying to @ollama

Model page: ollama.com/library/glm-5.2

And that's a wrap on World's Fair in SF 🎉 Thanks to everyone who stopped by the booth to say hi and dive into LanceDB internals. Some of the best conversations happen at a conference booth and this year was no exception. It was great meeting everyone who came out to SPIN SF

I tested Ollama's "90% faster Gemma 4" claim on an M5 Max MacBook Pro

Bottom line up front: Gemma 4 now runs on Ollama at a blazing 152 tok/s on an M5 Max MacBook Pro using the new MLX-optimized 26B MoE model. You should run it! "ollama pull gemma4:31b-mlx" and you're...

Congratulations to our friends at @togethercompute. Exciting moment for open models!!

Vipul Ved Prakash

@vipulved

Jul 1

We @togethercompute believe intelligence should be abundant, not expensive. Today we announced our Series C funding of $800m @ $8.3B valuation, to continue to build the world's most efficient platform for generative AI. Thanks @nikogallogly for telling our story in @nytimes!

Gemma 4 is now nearly 90% faster on Apple Silicon with Ollama using MLX! The speedup comes from improved multi-token prediction (MTP), now on by default for Gemma 4, with more models to come. Ollama automatically tunes how many tokens to draft as it runs, so it never slows

Read more:

Faster Gemma 4 on MLX with multi-token prediction· Ollama Blog

Run Ornith with Ollama: ollama run ornith For coding, use it with Claude or Pi: ollama launch claude --model ornith ollama launch pi --model ornith For the more capable 35B model, use: ollama launch claude --model ornith:35b

Ornith

@ornith_

Jun 25

Aloha! 🌺 Meet Ornith-1.0, a family of open-source LLMs specialized for agentic coding. Ornith-1.0 spans the full parameter sizes including 9B Dense, 31B Dense, 35B MoE, and 397B MoE. It achieves state-of-the-art performance among open-source models of comparable size on

Model page:

A self-improving family of open-source models for agentic coding

5.2 could be better with more RL ...

青龍聖者

@bdsqlsz

Jun 24

Deepswe's benchmark results are my own experience. I've used all models, GLM 5.2 ≈ Claude Opus 4.6–4.7. Kimi 2.7 code more like inference optimization. Looking forward to K3. Doubao-seed 2.1 Pro around 37% ≈ Gemini 3.5 Flash. code are quite weak, but visual are strong.

At @aiDotEngineer World's Fair next week? Come join us Tuesday night 🏓 We're co-hosting an evening at SPIN with @Theoryvc and @ollama. Talk shop with the people building the next-generation of local and cloud AI infrastructure, grab a drink, and get a few games in. 📅 Tuesday,

BYOK is now live in the GitHub Copilot App! Works with @ollama, foundry, and any OAI completions or Anthropic compatible messages endpoint. Give it a try today!

.@opencode and @ollama if you ask me

Suhail

@Suhail

Jun 22

Which means, somebody will make a lot of money commoditizing it vs charging a subscription for it.

The sharpest questions in AI live at the local–cloud boundary : where should inference run, & where should your data live? In town for AIE? Come hash it out with @Theoryvc, @lancedb & @ollama. June 30 · 6–9PM · SPIN SF 🏓