FlashML-org/FreeToken
FreeToken is an edge-native MoE inference engine for running large open-weight models on consumer hardware.
As of Nov 12, 2026, FlashML-org/FreeToken has 17,051 GitHub stars, +386 since Nov 5, 2026, #29 among Model & inference repos this week.
- Total stars
- 17,051
- Last day
- +32
- Since Nov 5, 2026
- +386
- Weekly rank
- #225
- Category
- Model & inference repos
- Language
- Python
Stars gained per day
Nov 5, 2026 – Nov 11, 2026 · peak +86/day
Links
FreeToken alternatives in Model & inference repos
- cactus-compute/cactus · 14,682 ★ · +8,502 /7dCactus is an edge inference engine with custom quantization, kernels and a graph runtime for mobile and embedded LLMs.
- NVIDIA/auto-ontology · 12,056 ★ · +5,917 /7dNVIDIA framework that builds semantic ontology layers over enterprise metadata to ground LLM and agent queries.
- tashfeenahmed/freellmapi · 50,937 ★ · +3,197 /7dAggregates free tiers from 34 LLM providers into one OpenAI-compatible endpoint with routing and failover.
- Niko1221/Strata · 43,375 ★ · +3,113 /7dStrata is a local inference engine running Qwen3.8-Flash-Next on consumer hardware.
- Alishahryar1/free-claude-code · 68,804 ★ · +2,260 /7dA proxy that routes Claude Code and other coding agents to 61 free and paid model providers.
- JustVugg/colibri · 50,091 ★ · +2,175 /7dA C inference engine that streams MoE experts from NVMe to run large models on limited hardware.
- cheaper-inference/OmniRoute · 86,707 ★ · +2,035 /7dOpen-source MIT AI gateway routing one endpoint to 375 providers and 1,400+ models.
- Zagitta/gpu-unlock · 18,823 ★ · +1,958 /7dPatches Linux to unlock full tensor-core throughput on GeForce RTX 20/30/40 GPUs.
Badge for your README
Data: daily star-total snapshots from the GitHub GraphQL API; stars gained = difference between snapshots. GitHub search and GH Archive are used to find repos. Updated daily around 02:30 UTC, counting through the last complete UTC day. · By Kai
How we rank · Weekly archive · RSS · JSON API