Top New Best Ask Show Jobs

OsamaJaber

268 karma · joined October 19, 2025

submissions comments

System Zero

rightnowai.co·2 pts·OsamaJaber·

Free Cached Input

runinfra.ai·3 pts·OsamaJaber·

Ornith 1.5 122B at 302tps

runinfra.ai·4 pts·OsamaJaber·

The fastest DeepSeek v4.1 flash 1017 – Day 0

runinfra.ai·3 pts·OsamaJaber·

Is this the best inference for agents?

runinfra.ai·4 pts·OsamaJaber·

DeepSeek v4.1 Flash at 378 tok/s 99.7% Cache hit rate

runinfra.ai·2 pts·OsamaJaber·

GLM 5.3 flash on AMD GPUs 670 tok/s

twitter.com·2 pts·OsamaJaber·

Agent loops stop re-reading your codebase on every turn

runinfra.ai·2 pts·OsamaJaber·

GLM 5.3 Flash faster and cheaper

runinfra.ai·1 pts·OsamaJaber·

The fastest and cheapest GLM 5.3 Flash endpoint

runinfra.ai·3 pts·OsamaJaber·

DeepSeek V4 Pro at 207 tok/s with the full 1M context, no quantization

runinfra.ai·3 pts·OsamaJaber·

Fastest Inference in MENA

runinfra.ai·2 pts·OsamaJaber·

DeepSeek V4 Flash at 278 tok/s, full precision, no quantization

runinfra.ai·10 pts·OsamaJaber·

The fastest full-precision Nemotron 3.5 Lightning endpoint (540 tok/s)

runinfra.ai·5 pts·OsamaJaber·

With software alone, one B200 beats the LPU and gets close to Cerebras

runinfra.ai·11 pts·OsamaJaber·

Kimi K3 2.78T on One CPU with 8GB RAM

github.com·4 pts·OsamaJaber·

Mixture-of-Kittens: An MoE training megakernel for NVL72

twitter.com·3 pts·OsamaJaber·

Lossless Inference

runinfra.ai·4 pts·OsamaJaber·

DeepSeek V4 Flash 2.98x faster, lossless

runinfra.ai·1 pts·OsamaJaber·

DeepSeek-V4-Flash 2.98x faster on 4x B200, lossless

twitter.com·3 pts·OsamaJaber·

What LLM Inference Costs

twitter.com·2 pts·OsamaJaber·

Kimi k3 run on RTX 5090

github.com·4 pts·OsamaJaber·

Kimi k3 now runs on one consumer GPU

twitter.com·8 pts·OsamaJaber·

The AGI Compiler "Auto"

github.com·2 pts·OsamaJaber·

Compiler for LLMs, world models, and AGI

arxiv.org·2 pts·OsamaJaber·

Auto: The AGI Compiler

github.com·4 pts·OsamaJaber·

Build You Own Model

runinfra.ai·2 pts·OsamaJaber·

RunInfra: Optimize any open model down to the kernel, deploy in 5 min

runinfra.ai·2 pts·OsamaJaber·

Compiles any HuggingFace model into a single persistent megakernel

twitter.com·2 pts·OsamaJaber·

Mega Kernels, Written by Agents

arxiv.org·2 pts·OsamaJaber·

Page 1 of 2Next →