The week in brief
Behind it, vLLM (+1,502) and Ollama (+1,081) followed.
Four new tools joined the directory this week: Strata, DeepSeek-V4.1-Flash, Recursive Flint and MOSS-Image (OpenMOSS).
On Hacker News, “WhatsApp's help pages hid LLM instructions: dont recommend a regulator” drew the most attention in this space (743 points, 222 comments). Among listed tools, Mistral AI's site was linked from 20 Hacker News stories over the past 90 days, more than any other tool here.
In research, “WeVisOne-VL: Unified Visual Perception from Objects to Pixels” collected the most upvotes on Hugging Face Daily Papers this week (244). Among new GitHub repositories created in the past 28 days, Dreamer-Toby/STEPQuant leads with 98 stars.
Biggest GitHub star gainers
GitHub star counts for the 52 open source tools we track, compared between Oct 18, 2026 and Oct 26, 2026 (180 hours apart).
| Tool | Stars before | Stars after | Gain | Change |
|---|---|---|---|---|
| ComfyUI | 138,563 | 143,105 | +4,542 | 3.28% |
| vLLM | 94,074 | 95,576 | +1,502 | 1.60% |
| Ollama | 183,742 | 184,823 | +1,081 | 0.59% |
| SGLang | 37,283 | 38,190 | +907 | 2.43% |
| llama.cpp | 131,627 | 132,396 | +769 | 0.58% |
| Whisper | 110,827 | 111,289 | +462 | 0.42% |
| Transformers | 167,822 | 168,182 | +360 | 0.21% |
| Unsloth | 78,229 | 78,580 | +351 | 0.45% |
New listings this week
| Tool | Category | Listed |
|---|---|---|
| Strata | ML Frameworks | Oct 24, 2026 |
| DeepSeek-V4.1-Flash | Open Weight Models | Oct 24, 2026 |
| Recursive Flint | Open Weight Models | Oct 23, 2026 |
| MOSS-Image (OpenMOSS) | Open Weight Models | Oct 22, 2026 |
Most discussed on Hacker News
The highest scoring Hacker News stories in this space during the week.
Listed tools with the most Hacker News links
Hacker News stories linking to each tool's own website over the past 90 days.
Notable new papers
The most upvoted papers on Hugging Face Daily Papers and the newest relevant arXiv preprints from this week.
- WeVisOne-VL: Unified Visual Perception from Objects to Pixels
- IQuest-Q1: Advancing Foundation Capabilities for Agentic CLI Systems
- Toward Recursive Self Improvement in Physical AI
- Self Distilling Future Video Representations for World Action Models
- FilmFlow: Agentic System for 3D Virtual Film Cinematography
- Measure Before You Train: Typed Decision Models vs. a Classical Baseline vs. a Generative LLM for Routing a Personal Multilingual Inbox
- Where Rounding Happens Matters: Rounding-Location Alignment Stabilizes LLM Reinforcement Learning
- Trusting the Tool: Asymmetric Deference of LLM Vulnerability Auditors to Static-Analysis Evidence
- Do "True" and "False" Add Up to One? Complement Consistency in LLM Agent Opinion Simulations
- J-Comm: Jacobian-Guided Latent Communication between LLM Agents
New GitHub repos in LLM research
Repositories created in the 28 days before Oct 26, 2026, ranked by stars.
| Repository | Stars | Language | Description |
|---|---|---|---|
| Dreamer-Toby/STEPQuant | 98 | Python | STEPQuant: When and Where Errors Matter in Delta-Rule Recurrent State Quantization |
| TheLastOxygen/mafdet | 56 | Python | You type. It already knows. Local LLMs that read and think while you type. |
| AllSpark-Research/Vela | 31 | Vela: User Behavior as Supervision Sparks LLM Intelligence (official paper, task templates, and prompts from AllSpark at Xiaohongshu). |
How we compile this report
Every figure comes from a public source: star counts from the GitHub API, story scores and links from Hacker News (via Algolia), papers from arXiv and Hugging Face Daily Papers, and new repositories from GitHub search. Hacker News link counts include stories whose link points to the tool's own domain. Listing dates are the day a tool joined Axiomi. Numbers are shown as recorded; if a figure is unavailable, it is left out rather than estimated.