llamaperf

Kolibri-1

1 report

Kolibri-1 78B · 3.5B active at 2-bit generates a median of 27.3 tokens per second on the AMD Radeon 890M, the hardware it is most run on, from one plain run (community reports, as of 21 Nov 2026).

As of 21 Nov 2026, Kolibri-1 78B · 3.5B active at 2-bit on the hardware it is most run on, with the median of plain runs (one device, one request, no speculative decoding, the whole model in its memory):

Kolibri-1 VRAM requirements by size and quant →

How does Kolibri-1 run on your hardware?

Pick your GPU or Mac, then read the speeds people reported for Kolibri-1 on it, or estimate memory fit and speed.

Free to use. No account needed. Memory estimates and community measurements are labelled separately.

Run Kolibri-1 yourself? Add your numbers.

Your coding agent can run a speed test on your machine and send it here with a free account. Or paste a result you've already posted.

Reported performance for Kolibri-1

Filter this model’s reports by setup →
Thin page (1 of 3 reports needed for indexing). Add yours.
Tone: mixedAMD hardware
generation:
27.3 tokens/s
prompt processing (prefill):
430.0 tokens/s
quant:
IQ2_M (GGUF)

Reported by the source; GPU count, offloading and concurrent requests can change these figures. Prompt-processing speed is input throughput, not time to first token. Check the full setup before comparing.

agenticcoding

User benchmarks Kolibri-1 78B at 27.28 t/s decode on a Minisforum AI X1 Pro with a Radeon 890M and 96GB RAM (48GB VRAM share). Setup is llama.cpp over Vulkan with IQ2_M at 262k context; prefill is 430 t/s and the model scored 5/7 on the user's gauntlet. User ranks it the fastest smart agent on the box, ahead of Qwen3.6-35B at 26.8-27.0 t/s and Qwen3-Next-80B at 26 t/s, and notes nothing bigger clears 10 t/s except small-active MoEs.

Nov 20, 2026

Get a weekly email of new Kolibri-1 reports on any GPU.

Email me new reports

Reported configurations

Individual observations from this page, not expected speeds or a ranking. Select a model to open its report, with the source and offloading conditions. Reported context may be a limit rather than actual input length.

Model / hardwareQuant / engineReported contextGeneration
Kolibri-1 78B (3.46B active) A3.5B
AMD Radeon 890M
IQ2_M
llama.cpp
262,00027.3 tokens/s

Frequently asked

What hardware do people run Kolibri-1 on?

The hardware Kolibri-1 78B · 3.5B active at 2-bit is most run on: AMD Radeon 890M (27.3 t/s median, one run). Medians of plain runs, as of 21 Nov 2026.

Which inference engines do people use for Kolibri-1?

Reports of Kolibri-1 name llama.cpp (1 report), as of 21 Nov 2026.