Georgi Gerganov ✓

@ggerganov · joined 27 May 2015

24th at the Electrica puzzle challenge | building https://t.co/baTQS2bL7I | engineer @huggingface

71 426Followers

294Following

1 771Posts total

1.6MViews on collected posts

Against accounts of the same size

14 posts from the last 90 days, next to the 10K–100K follower range. shown widely, but few of those viewers react.

Median views28 044this account942median for 10K–100K

Reach, %39.26%this account3.24%median for 10K–100K

Engagement, %0.58%this account1.99%median for 10K–100K

Metric	This account	Median for 10K–100K	Ratio
Median views per post	28 044	942	29.8×
Reach (views ÷ followers)	39.26%	3.24%	12.1×
Engagement rate	0.58%	1.99%	0.29×

Others in this range → Compare with another account → How these benchmarks are built →

Growth & engagement

How the posts we collected actually performed: views and reaction rate post by post, what the audience did with them, and where the follower count goes.

Views per post

459.2K14 Aug

32.1K

32.5K

22.1K

24K

14K

60.3K17 Aug

110.2K

742.3K

5.4K

8.1K

4.6K18 Aug

19.5K1 Sep

46.1K4 Sep

Last 14 collected posts, oldest on the left. The scale is logarithmic: one post can outrun the rest a hundred times over.

Engagement rate per post

0.47%14 Aug

0.58%

0.82%

0.24%

0.58%

0.52%

0.40%17 Aug

0.51%

0.44%

0.74%

1.16%

0.99%18 Aug

0.67%1 Sep

1.36%4 Sep

Reactions — likes, reposts, replies and quotes — divided by views. Median for 10K–100K accounts is 1.99%.

What the audience does

Likes58.0%6 942 in total

Reposts4.3%511 in total

Replies2.2%260 in total

Quotes1.2%143 in total

Bookmarks34.4%4 115 in total

Share of every reaction we collected for this account. Replies mean argument, reposts mean endorsement, bookmarks mean the post was worth keeping.

The follower curve appears once this account has two daily snapshots — we take one a day, and this one is on its first.

Latest posts

Hugging Face has been acquired by NVIDIA It is quite exciting to be a part of this journey! NVIDIA has been an active supporter of the llama.cpp project. For more than a year now, their engineers have actively contributed to the codebase, collaborated with the community and 46.1K views · 542 likes · 48 reposts · 25 replies 04 Sep 2026 · Open on X →

The Llama app for Mac now comes with a simple request builder for llama.cpp's REST API https://t.co/irwUZEykBC 19.5K views · 111 likes · 6 reposts · 13 replies 01 Sep 2026 · Open on X →

@ggerganov but... https://t.co/rG7tatuHFA 4.6K views · 37 likes · 1 reposts · 7 replies 18 Aug 2026 · Open on X →

@ggerganov what if opus5 has been a thick 27B all along...? https://t.co/HCf2Q0xXJy 8.1K views · 90 likes · 1 reposts · 3 replies 17 Aug 2026 · Open on X →

@ggerganov I really do hope you're not posting that with straight face 5.4K views · 38 likes · 0 reposts · 2 replies 17 Aug 2026 · Open on X →

let that sink in https://t.co/rCficQYZB7 742.3K views · 2.8K likes · 249 reposts · 120 replies 17 Aug 2026 · Open on X →

I find the "inception" pattern to be very useful in many agentic use cases. You can force the model to take an action when it thinks for too long by injecting a thought after a specified reasoning budget. Helps dealing with underspecified tasks which make the model reason for 110.2K views · 488 likes · 36 reposts · 27 replies 17 Aug 2026 · Open on X →

to limit the max reasoning length, add: ... \ --reasoning-budget 4096 \ --reasoning-budget-message "... I am thinking for too -- let me gather more info about the task." adjust to your needs 60.3K views · 216 likes · 13 reposts · 7 replies 17 Aug 2026 · Open on X →

@ggerganov 16GB RTX 5080 Laptop: llama serve \ -hf unsloth/Qwen3.8-27B-GGUF \ -hff Qwen3.8-27B-UD-IQ3_XXS.gguf \ --no-mmproj -c 65536 \ -ctk q8_0 -ctv q8_0 \ -b 1024 -np 1 \ --spec-default --spec-type draft-mtp \ --reasoning-preserve --fit off --agent ~60 t/s. 14K views · 65 likes · 4 reposts · 2 replies 14 Aug 2026 · Open on X →

on DGX Spark: llama serve \ -hf ggml-org/Qwen3.8-27B-GGUF:Q4_K_M \ -hfd ggml-org/Qwen3.8-27B-GGUF:Q4_0 \ --spec-default \ --spec-type draft-mtp \ --reasoning-preserve --agent 24K views · 132 likes · 3 reposts · 5 replies 14 Aug 2026 · Open on X →

enjoy! 22.1K views · 49 likes · 2 reposts · 3 replies 14 Aug 2026 · Open on X →

32GB VRAM (e.g. RTX 5090): llama serve \ -hf ggml-org/Qwen3.8-27B-GGUF:Q4_K_M \ -hfd ggml-org/Qwen3.8-27B-GGUF:Q4_0 \ --spec-default \ --spec-type draft-mtp \ --ctx-size 196608 \ --cache-type-k q8_0 \ --cache-type-v q8_0 \ --reasoning-preserve --fit off --agent 32.5K views · 242 likes · 12 reposts · 9 replies 14 Aug 2026 · Open on X →

advanced: llama serve \ -hf ggml-org/Qwen3.8-27B-GGUF \ --spec-default \ --spec-type draft-mtp \ --reasoning-preserve --agent 32.1K views · 180 likes · 4 reposts · 3 replies 14 Aug 2026 · Open on X →

simple: llama serve -hf ggml-org/Qwen3.8-27B-GGUF --spec-type draft-mtp 459.2K views · 2K likes · 132 reposts · 34 replies 14 Aug 2026 · Open on X →