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

Re:VibeThinker: 3B param model that beats Opus 4.5 on reason... by nsingh2 on Tuesday June 23, 2026 @12:52AM (#48640468)

Lots of confusion about what this model is actually focused on.

It is a cheap specialist for closed-world, verifiable reasoning tasks like math, self-contained coding problems, and similar.

"Closed-world" means the needed information is already in the context. It is not a tool-using agent that can discover missing context. "Verifiable" means answers are hard to generate but easy to check.

So no open ended research, repo wide agent work, factual Q&A, or SVG generation. More of a compact reasoning module for bounded problems.

[ Parent | on HN ]

Re:VibeThinker: 3B param model that beats Opus 4.5 on reason... by nsingh2 on Tuesday June 23, 2026 @02:13AM (#48641031)

To follow up on this, I had it solve a nasty ODE problem that I saw in the recent Mathematica 15 release post:

    Solve the following first-order ODE for f(x):

    ((-1 - 2*x)*f(x)*tan(1 + x - exp(-61 - 2*x)*f(x)/x)
    + exp(61 + 2*x)*x*(1 - x*tan(1 + x - exp(-61 - 2*x)*f(x)/x))
    + x*tan(1 + x - exp(-61 - 2*x)*f(x)/x)*f'(x)) = 0

    Find the general solution f(x).

And surprisingly it found a valid solution! Extra impressive because it runs 25 tok/s on my measly RTX 2070 super.

    f(x) = x*exp(61 + 2*x)*(1 + x - arccos(C/x))

    C is an arbitrary constant.

Apparently Mathematica 14.3 couldn't solve this ODE.

[ Parent | on HN ]

Re:VibeThinker: 3B param model that beats Opus 4.5 on reason... by trick-or-treat on Tuesday June 23, 2026 @03:30AM (#48641549)

How do we know the solution isn't in the weights though?

[ Parent | on HN ]

Re:VibeThinker: 3B param model that beats Opus 4.5 on reason... by le-mark on Tuesday June 23, 2026 @06:49AM (#48643021)

How do you know it’s a valid solution? Are you able to verify it yourself?

[ Parent | on HN ]

Re:VibeThinker: 3B param model that beats Opus 4.5 on reason... by krageon on Tuesday June 23, 2026 @09:48AM (#48644947)

This is a math problem with a math solution. You verify it with math

[ Parent | on HN ]

Re:VibeThinker: 3B param model that beats Opus 4.5 on reason... by kanbankaren on Tuesday June 23, 2026 @08:58AM (#48644291)

Qwen 3.5 9B Q4_K_M solved this using 10K tokens in 5 mins on a RX 7600.

The answer is exactly what you have posted. I am impressed by Qwen!

[ Parent | on HN ]

Re:VibeThinker: 3B param model that beats Opus 4.5 on reason... by skeledrew on Tuesday June 23, 2026 @04:12AM (#48641886)

If it can code well then once you put it in a loop with an interpreter it can do anything.

[ Parent | on HN ]