kaggle-tpu-lab

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.

ARahim3/kaggle-tpu-lab

View on GitHub ↗

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.

kaggle-tpu-lab grew +15.5% in the last 30 days, ranks 106th of 397 in Production Serving & Deployment, last commit 46 days ago.

Star trajectory

71 days tracked
2026-09-11813 stars today2026-11-22
Momentum

+15.5% in 30d, above the Production Serving & Deployment median of +1.3%.

Maintenance

Last commit 46 days ago, staler than the median of 3d.

How it compares in Production Serving & Deployment

kaggle-tpu-lab
813
stars
+15.5% 30d
commit 46 days ago
Production Serving & Deployment median
272
stars
+1.3% 30d
commit ~3d ago
Leader · vllm
97.3k
stars
+2.2% 30d
commit today

Ranked 106th of 397 by stars.

Ranked by shared tags, then keywords, then stars
  1. The fastest inference engine for Qwen3.8-27B on the NVIDIA RTX 5090: up to 500 tokens/s for one agent and up…

    Committed 17d ago · MIT

    Shares Production Serving & Deployment, Inference Optimization

  2. Run large AI models on your own PC: Qwen3.8-Flash-Next on Strata's engine, and GLM 5.3 Flash, DeepSeek V4…

    +962 (+243%) in 30d · Committed 3d ago · C++ · MIT

    Shares Inference Optimization, Production Serving & Deployment

  3. Xolotl: run DeepSeek-V4.1-Flash, all 552B, as a local coding agent on one gaming PC (RTX 4090 + 110 GB RAM)…

    +264 (+90%) in 30d · Committed today · C++ · MIT

    Shares Production Serving & Deployment, Inference Optimization

  4. Inference engine for Intel devices. Serve LLMs, VLMs, Whisper, Kokoro-TTS, Embedding and Rerank models over…

    +10 (+1.8%) in 30d · Committed yesterday · Python · Apache-2.0

    Shares Production Serving & Deployment, Inference Optimization

  5. llama.cpp/ik_llama.cpp launcher: loads big MoE models across mismatched multi-GPU rigs by exact VRAM math.

    +2 (+0.7%) in 30d · Committed 2d ago · Go · MIT

    Shares Inference Optimization, Production Serving & Deployment

  6. Alzeo is an open-source installer that turns an NVIDIA DGX Spark, or any GB10 machine (ASUS, Dell, HP, Lenovo…

    +36 (+15%) in 30d · Committed 2d ago · Python · MIT

    Shares Production Serving & Deployment, Inference Optimization

  7. run a 27b local ai on your gaming gpu, rtx 3060 and up, radeon rx 6600 and up in beta, 8gb or more, linux and…

    +31 (+19%) in 30d · Committed 35d ago · Python

    Shares Production Serving & Deployment, Inference Optimization

  8. An inference framework customized for AI MAX 395, compatible with both Linux and Windows, and supports…

    +45 (+41%) in 30d · Committed today · C++ · AGPL-3.0

    Shares Inference Optimization, Production Serving & Deployment

  9. Bonsai 2 27B: 262k context (trained max) on 12 GB, 54 to 97 tok/s decode, 2x prefill, q8_0 KV recipe, grammar…

    Committed yesterday · Python · Other license

    Shares Inference Optimization, Production Serving & Deployment

  10. Local proxy that learns your app's typed LLM decisions and answers them with a Laya head. Jev and OpenAI…

    +3 (+4.1%) in 30d · Committed 41d ago · Python · Apache-2.0

    Shares Inference Optimization, Production Serving & Deployment

Language
Python
License
MIT License
Contributors
1
Forks
110
Watchers
7
Open issues
18
Releases
0
Created
2026-08-31
Stars this week
+12

You've seen where kaggle-tpu-lab stands today.

Get the moves before everyone else: the 5 Production Serving & Deployment repos climbing fastest, every Tuesday.

Free weekly digest. Double opt-in, unsubscribe anytime.

Maintain kaggle-tpu-lab?

Claim this page to add your own call-to-action.

Claim this page