kaggle-tpu-lab
by ARahim3 (personal account)
Frontier-class open models on a free Kaggle TPU v5e-8: GLM-5.3-Flash 320B MoE (~165 tok/s, our own JAX engine) and Qwen3.8-27B bf16 (~130 tok/s), 262k context, prefix caching. Works with Claude Code, Codex, opencode and pi. (description from the project’s GitHub page)
Repository facts
- Stars
- 679
- Forks
- 93
- Open issues and pull requests
- 15
- Primary language
- Python
- License
- MIT License (MIT)
- Owner
- ARahim3 · Personal account
- Created
- Aug 31, 2026
- Last push
- Oct 6, 2026
Topics
Related projects
Projects that share this one’s GitHub topics.
Rank 6,051: peonist-ai/halogen-flash-server
The fastest way to run Qwen3.8-Flash-Next on Strix Halo (gfx1151)
1,018 starsShellOther licenseLast push Oct 18, 2026
Rank 7,564: pegainfer-project/pegainfer
Pure Rust + CUDA LLM inference engine — no PyTorch, OpenAI-compatible, serves Qwen3 to Kimi-K2
724 starsRustApache-2.0Last push Oct 17, 2026
Rank 3,585: syv-ai/HyperQwen
Serve large Qwen models fast on the GPUs you actually own. Qwen3.8-27B on a single 24 GB card with vLLM: 127 tok/s single-user (381 when the answer quotes the prompt), ~1,035 tok/…
2,054 starsPythonApache-2.0Last push Oct 18, 2026
Rank 1,979: raullenchai/Rapid-MLX
Rapid-MLX is an open-source (Apache 2.0) OpenAI- and Anthropic-compatible LLM inference server for Apple Silicon, built on MLX, focused on reliable tool calling for coding agents.…
3,996 starsPythonOther licenseLast push Oct 18, 2026
Rank 43: vllm-project/vllm
A high-throughput and memory-efficient inference and serving engine for LLMs
94,068 starsPythonApache-2.0Last push Oct 18, 2026
Rank 1,389: gpustack/gpustack
A GPU cluster manager for high-performance AI model serving (vLLM, SGLang) and on-demand SSH-accessible GPU instances.
5,844 starsPythonApache-2.0Last push Oct 16, 2026