Dark | Light
[GUEST ACCESS MODE: Data is scrambled or limited to provide examples. Make requests using your API key to unlock full data. Check https://lunarcrush.ai/auth for authentication information.]

[@reach_vb](/creator/twitter/reach_vb)
"BOOM: We've just re-launched HuggingChat v2 πŸ’¬ - XXX open source models in a single interface is stronger than ChatGPT πŸ”₯ Introducing: HuggingChat Omni πŸ’« Select the best model for every prompt automatically πŸš€ Automatic model selection for your queries XXX models available across XX providers including @GroqInc @CerebrasSystems @togethercompute @novita_labs and more Powered by HF Inference Providers access hundreds of AI models using only world-class inference providers Omni uses a policy-based approach to model selection (after experimenting with different methods). Credits to @katanemo_"  
[X Link](https://x.com/reach_vb/status/1978854312647307426) [@reach_vb](/creator/x/reach_vb) 2025-10-16T16:03Z 33.6K followers, 61.6K engagements


"Kudos to OpenAI team for their guide on what makes documentation good - it's brilliant πŸ₯³ Few gems that stood out: Make docs easy to skim clear titles short paras bullets & bold text Write simply cut jargon avoid abbreviations explain even obvious steps Be broadly helpful docs arent just for experts theyre for everyone Prioritise value solve common problems first avoid teaching bad habits Above all: docs are an act of empathy ❀ If you care about Developer Experience you should give it a read"  
[X Link](https://x.com/reach_vb/status/1980309939961602550) [@reach_vb](/creator/x/reach_vb) 2025-10-20T16:27Z 33.6K followers, 51.3K engagements


"Pretty cool to see a MIT licensed 15B model competing w/ DeepSeek R1 - how are the vibes πŸ‘€"  
[X Link](https://x.com/reach_vb/status/1973731570008871381) [@reach_vb](/creator/x/reach_vb) 2025-10-02T12:47Z 33.5K followers, 33.3K engagements


"@karpathy Love itt btw @kszucs_ from the team opened a PR to download the files faster and efficiently from Hugging Face πŸ€—"  
[X Link](https://x.com/reach_vb/status/1978062616976044540) [@reach_vb](/creator/x/reach_vb) 2025-10-14T11:37Z 33.5K followers, 1319 engagements


"NEW: GLM-4.5 & GLM-4.5-Air from @Zai_org - competitive w/ claude X opus and beats Gemini XXX Pro MIT licenseπŸ”₯ GLM-4.5: 355B total params 32B active (MoE) GLM-4.5-Air: 106B total params 12B active (MoE) "Thinking mode" (complex tasks) + "Non-thinking mode" (instant responses) 128K context length + native function calling Impressive benchmarks: AIME24: XXXX (vs. Claude X Opuss 75.7) MATH 500: XXXX (vs. GPT-4.1s 96.7) GPQA: XXXX (vs. Gemini XXX Pros 84.4) SWE-bench Verified: XXXX (vs. Claude X Sonnets 70.4) Terminal-Bench: XXXX (vs. Claude X Opuss 43.2) MoE - Loss-free balance routing + sigmoid"  
[X Link](https://x.com/reach_vb/status/1949823507400884446) [@reach_vb](/creator/x/reach_vb) 2025-07-28T13:25Z 33.5K followers, 52.4K engagements


"@SiddhantSadangi But you can just use the models via chat completion/ responses directly no"  
[X Link](https://x.com/reach_vb/status/1978948267934175651) [@reach_vb](/creator/x/reach_vb) 2025-10-16T22:16Z 33.5K followers, XX engagements


"Pretty wild what 900M params can do - PaddleOCR VL πŸ”₯ SOTA on OmniDocBench v1.0 & v1.5 (text tables formulas charts reading order) Multilingual - XXX languages (Latin Arabic Cyrillic Devanagari Thai etc) Handles handwriting historical docs noisy scans Supports output JSON + Markdown NaViT-style dynamic visual encoder + ERNIE-4.5-0.3B LM Available on HF and works w/ Transformers (remote code) πŸ€—"  
[X Link](https://x.com/reach_vb/status/1979219167258554752) [@reach_vb](/creator/x/reach_vb) 2025-10-17T16:13Z 33.6K followers, 15.9K engagements


"Letsss gooo DeepSeek just released a 3B OCR model on Hugging Face πŸ”₯ Optimised to be token efficient AND scale 200K+ pages/day on A100-40G Same arch as DeepSeek VL2 Use it with Transformers vLLM and more πŸ€—"  
[X Link](https://x.com/reach_vb/status/1980170192392270227) [@reach_vb](/creator/x/reach_vb) 2025-10-20T07:12Z 33.6K followers, 108.8K engagements


"You can read the full text here:"  
[X Link](https://x.com/reach_vb/status/1980310081511288918) [@reach_vb](/creator/x/reach_vb) 2025-10-20T16:28Z 33.6K followers, 2639 engagements


"@JulienBlanchon @pontusab still need to find time to try out svelte πŸ˜…"  
[X Link](https://x.com/reach_vb/status/1980604348380983481) [@reach_vb](/creator/x/reach_vb) 2025-10-21T11:57Z 33.6K followers, XX engagements


"The Hugging Face Hub team is on a tear recently: You can create custom apps with domains on spaces Edit GGUF metadata on the Fly XXX% of the Hub is powered by Xet - faster efficient Responses API support for ALL Inference Providers MCP-UI support for HF MCP Server Search papers based on the Org Showcase repository size on the UI and a lot more - excited for the coming weeks/ months as we continue to improve the overall UX πŸ€—"  
[X Link](https://x.com/reach_vb/status/1976328324503744938) [@reach_vb](/creator/x/reach_vb) 2025-10-09T16:46Z 33.5K followers, 14.3K engagements


"chat is this real 128K context int4 quantisation 1B params distilled from Llama X πŸ”₯"  
[X Link](https://x.com/reach_vb/status/1978911810494345401) [@reach_vb](/creator/x/reach_vb) 2025-10-16T19:51Z 33.5K followers, 32.1K engagements


"@xlr8harder @rishdotblog very cool happy to throw in some credits via HF Inference Providers if you want to take it out for a spin :D"  
[X Link](https://x.com/reach_vb/status/1980281182475428158) [@reach_vb](/creator/x/reach_vb) 2025-10-20T14:33Z 33.6K followers, XX engagements


"🚨 Apple just released FastVLM on Hugging Face - XXX XXX and 7B real-time VLMs with WebGPU support 🀯 85x faster and 3.4x smaller than comparable sized VLMs 7.9x faster TTFT for larger models designed to output fewer output tokens and reduce encoding time for high resolution images Bonus: works in REALTIME directly in your browser powered by transformers.js and WebGPU πŸ”₯ Try it out on the demo below πŸ‘‡"  
[X Link](https://x.com/reach_vb/status/1961471154197053769) [@reach_vb](/creator/x/reach_vb) 2025-08-29T16:49Z 33.6K followers, 585K engagements


"Oh wow Qwen dropped smol 2B VLM πŸ”₯"  
[X Link](https://x.com/reach_vb/status/1980670737611649356) [@reach_vb](/creator/x/reach_vb) 2025-10-21T16:21Z 33.6K followers, 16.9K engagements

[GUEST ACCESS MODE: Data is scrambled or limited to provide examples. Make requests using your API key to unlock full data. Check https://lunarcrush.ai/auth for authentication information.]

@reach_vb "BOOM: We've just re-launched HuggingChat v2 πŸ’¬ - XXX open source models in a single interface is stronger than ChatGPT πŸ”₯ Introducing: HuggingChat Omni πŸ’« Select the best model for every prompt automatically πŸš€ Automatic model selection for your queries XXX models available across XX providers including @GroqInc @CerebrasSystems @togethercompute @novita_labs and more Powered by HF Inference Providers access hundreds of AI models using only world-class inference providers Omni uses a policy-based approach to model selection (after experimenting with different methods). Credits to @katanemo_"
X Link @reach_vb 2025-10-16T16:03Z 33.6K followers, 61.6K engagements

"Kudos to OpenAI team for their guide on what makes documentation good - it's brilliant πŸ₯³ Few gems that stood out: Make docs easy to skim clear titles short paras bullets & bold text Write simply cut jargon avoid abbreviations explain even obvious steps Be broadly helpful docs arent just for experts theyre for everyone Prioritise value solve common problems first avoid teaching bad habits Above all: docs are an act of empathy ❀ If you care about Developer Experience you should give it a read"
X Link @reach_vb 2025-10-20T16:27Z 33.6K followers, 51.3K engagements

"Pretty cool to see a MIT licensed 15B model competing w/ DeepSeek R1 - how are the vibes πŸ‘€"
X Link @reach_vb 2025-10-02T12:47Z 33.5K followers, 33.3K engagements

"@karpathy Love itt btw @kszucs_ from the team opened a PR to download the files faster and efficiently from Hugging Face πŸ€—"
X Link @reach_vb 2025-10-14T11:37Z 33.5K followers, 1319 engagements

"NEW: GLM-4.5 & GLM-4.5-Air from @Zai_org - competitive w/ claude X opus and beats Gemini XXX Pro MIT licenseπŸ”₯ GLM-4.5: 355B total params 32B active (MoE) GLM-4.5-Air: 106B total params 12B active (MoE) "Thinking mode" (complex tasks) + "Non-thinking mode" (instant responses) 128K context length + native function calling Impressive benchmarks: AIME24: XXXX (vs. Claude X Opuss 75.7) MATH 500: XXXX (vs. GPT-4.1s 96.7) GPQA: XXXX (vs. Gemini XXX Pros 84.4) SWE-bench Verified: XXXX (vs. Claude X Sonnets 70.4) Terminal-Bench: XXXX (vs. Claude X Opuss 43.2) MoE - Loss-free balance routing + sigmoid"
X Link @reach_vb 2025-07-28T13:25Z 33.5K followers, 52.4K engagements

"@SiddhantSadangi But you can just use the models via chat completion/ responses directly no"
X Link @reach_vb 2025-10-16T22:16Z 33.5K followers, XX engagements

"Pretty wild what 900M params can do - PaddleOCR VL πŸ”₯ SOTA on OmniDocBench v1.0 & v1.5 (text tables formulas charts reading order) Multilingual - XXX languages (Latin Arabic Cyrillic Devanagari Thai etc) Handles handwriting historical docs noisy scans Supports output JSON + Markdown NaViT-style dynamic visual encoder + ERNIE-4.5-0.3B LM Available on HF and works w/ Transformers (remote code) πŸ€—"
X Link @reach_vb 2025-10-17T16:13Z 33.6K followers, 15.9K engagements

"Letsss gooo DeepSeek just released a 3B OCR model on Hugging Face πŸ”₯ Optimised to be token efficient AND scale 200K+ pages/day on A100-40G Same arch as DeepSeek VL2 Use it with Transformers vLLM and more πŸ€—"
X Link @reach_vb 2025-10-20T07:12Z 33.6K followers, 108.8K engagements

"You can read the full text here:"
X Link @reach_vb 2025-10-20T16:28Z 33.6K followers, 2639 engagements

"@JulienBlanchon @pontusab still need to find time to try out svelte πŸ˜…"
X Link @reach_vb 2025-10-21T11:57Z 33.6K followers, XX engagements

"The Hugging Face Hub team is on a tear recently: You can create custom apps with domains on spaces Edit GGUF metadata on the Fly XXX% of the Hub is powered by Xet - faster efficient Responses API support for ALL Inference Providers MCP-UI support for HF MCP Server Search papers based on the Org Showcase repository size on the UI and a lot more - excited for the coming weeks/ months as we continue to improve the overall UX πŸ€—"
X Link @reach_vb 2025-10-09T16:46Z 33.5K followers, 14.3K engagements

"chat is this real 128K context int4 quantisation 1B params distilled from Llama X πŸ”₯"
X Link @reach_vb 2025-10-16T19:51Z 33.5K followers, 32.1K engagements

"@xlr8harder @rishdotblog very cool happy to throw in some credits via HF Inference Providers if you want to take it out for a spin :D"
X Link @reach_vb 2025-10-20T14:33Z 33.6K followers, XX engagements

"🚨 Apple just released FastVLM on Hugging Face - XXX XXX and 7B real-time VLMs with WebGPU support 🀯 85x faster and 3.4x smaller than comparable sized VLMs 7.9x faster TTFT for larger models designed to output fewer output tokens and reduce encoding time for high resolution images Bonus: works in REALTIME directly in your browser powered by transformers.js and WebGPU πŸ”₯ Try it out on the demo below πŸ‘‡"
X Link @reach_vb 2025-08-29T16:49Z 33.6K followers, 585K engagements

"Oh wow Qwen dropped smol 2B VLM πŸ”₯"
X Link @reach_vb 2025-10-21T16:21Z 33.6K followers, 16.9K engagements

creator/twitter::874987512850128897/posts
/creator/twitter::874987512850128897/posts