k
keynews.ai
keynews.ai
FrontierThe story, in brief

VibeThinker: 3B param model that beats Opus 4.5 on reasoning with novel SFT+GRPO

3B parameters. Beats Claude Opus 4.5 on reasoning. Here's how VibeThinker did it.

Reporting by Hacker News 1 min brief
Illustration of a transparent lens revealing connected networks across layers of paper.
Exploring the next frontier of AI research.AI illustration by KeyNews
The KeyNews take

Why it matters

A smaller, open model claims superior reasoning performance on established benchmarks using novel training techniques (SFT+GRPO), challenging the narrative that scale and closed models dominate frontier reasoning capabilities.

The key facts

11 to know

  1. VibeThinker: 3B parameter model

  2. Claims to beat Anthropic Claude Opus 4.5 on reasoning benchmarks

  3. Training approach: SFT (Supervised Fine-Tuning) + GRPO (novel training method)

  4. Published on arXiv (academic venue, not peer-reviewed yet)

  5. Low engagement: 10 HN points, 0 comments as of publication

  6. UNVERIFIED: Claims not independently validated; benchmark details not specified in summary

  7. Model: VibeThinker, 3B parameters

  8. Benchmark claim: Beats Opus 4.5 on reasoning tasks

  9. Source: arXiv preprint (2606.16140), not peer-reviewed or industry-validated

  10. Community signal: Low engagement (10 HN points, 0 comments) suggests limited validation

  11. UNVERIFIED: No corroborating benchmarks, no comparison methodology details provided in headline

Go to the source

Hacker Newsarxiv.org

Publisher excerpt: Article URL: Comments URL: Points: 10 # Comments: 0
Read original report
Back to today's editionMore frontier news

The wider picture

View all
Illustration of a transparent lens revealing connected networks across layers of paper.
AI illustration by KeyNews
Frontier01

Alibaba Unveils Zhenwu V900 — and Plans Qwen Models With Up to 10 Trillion Parameters

Alibaba is advancing on two fronts simultaneously: announcing a custom AI accelerator (Zhenwu V900) and committing to massive model scale (10T parameters for future Qwen releases). For practitioners, this matters as a credible third-party capability play outside the US-China licensing squeeze; for enthusiasts, it's a significant lab-race signal about training compute and parameter scaling as competitive levers.

TechRepublic
Illustration of a transparent lens revealing connected networks across layers of paper.
AI illustration by KeyNews
Frontier02

Kyutai Releases Voice of Reason: A Speech-Native Model that Solves Spoken Math with Reinforcement Learning

A new capability frontier: models that reason directly in speech without transcription bottlenecks. This changes how we think about multimodal reasoning and what's possible with open-weight releases at scale.

MarkTechPost
Illustration of a transparent lens revealing connected networks across layers of paper.
AI illustration by KeyNews
Frontier03

Claude Opus 5.5, GPT-6 Sol, GPT-6 Luna, and a new price war

Frontier labs are shipping upgraded reasoning and multimodal models in rapid succession, signaling acceleration in the capability race. Simultaneous price cuts reshape AI economics for practitioners.

Simon Willison