The official API server for Exllama. OAI compatible, lightweight, and fast.
active 2024-03-17 → 2026-08-08 (UTC)
Activity over time
Daily event counts in the loaded window
Line chart, 875 days from 2024-03-17 to 2026-08-08. Pushes: 377 total, peak 8 in a day. Pull requests: 168 total, peak 5 in a day. Issues: 199 total, peak 4 in a day. Comments: 489 total, peak 30 in a day. Stars: 494 total, peak 8 in a day.
- Pushes
- Pull requests
- Issues
- Comments
- Stars
Stars, PRs, issues and forks are under-captured in the later part of this window. GH Archive progressively stopped capturing non-push events during 2026 — −95% or worse by the end of the window. Every series here except Pushes fades for that reason, so a decline above reflects the archive, not this repository. Pushes stay reliable throughout, so read them, and the contributor counts derived from them, as the real signal. Data health has the measurements.
Top contributors
Pushes, PRs, issues, reviews and comments — stars and forks excluded, so this is contribution rather than popularity
| Contributor | Contributions | Pushes | PRs | Comments |
|---|---|---|---|---|
| bdashore3 | 434 | 225 | 49 | 101 |
| kingbri1 | 203 | 96 | 27 | 56 |
| turboderp | 96 | 56 | 11 | 25 |
| DocShotgun | 70 | 0 | 19 | 51 |
| AlpinDale | 39 | 0 | 7 | 32 |
| Originalimoc | 31 | 0 | 3 | 20 |
| Orion-zhen | 17 | 0 | 2 | 11 |
| SecretiveShell | 13 | 0 | 10 | 3 |
| anunknowperson | 13 | 0 | 0 | 9 |
| lucyknada | 12 | 0 | 2 | 9 |
| strikeoncmputrz | 12 | 0 | 2 | 7 |
| bold84 | 11 | 0 | 4 | 6 |
| quarterturn | 9 | 0 | 0 | 9 |
| gittb | 9 | 0 | 3 | 6 |
| atisharma | 8 | 0 | 0 | 4 |
| official-elinas | 8 | 0 | 0 | 6 |
| TyraVex | 7 | 0 | 0 | 5 |
| frenzybiscuit | 6 | 0 | 0 | 5 |
| kreolsky | 6 | 0 | 0 | 2 |
| cikkle | 6 | 0 | 0 | 3 |
Recent activity
Latest issues, pull requests and releases
- Issue comment#430ethanj8012026-07-01 15:11Fix chat completions returning null logprobs for models without reasoning
- Issue comment#426SaSteMi2026-06-13 20:53[BUG] common.errors module missing but imported in multiple core files (main branch)
- Pull request#427SaSteMi2026-06-13 20:52
- Pull request#425turboderp2026-06-12 23:39
- Issue comment#417c6083452026-04-16 02:34[REQUEST] Please support CUDA 13
- Pull request#418RodriMora2026-04-12 19:06
- Issue comment#413turboderp2026-03-30 22:48Full tool-calling support, inference abort fixes, XML parsing, OpenAI streaming compliance
- Issue#416FQDianYan2026-03-16 07:38[BUG]Can't initialize tabbyAPI with Python 3.12
- Issue comment#415lesj06102026-02-25 12:01feat: add vLLM 0.16 mistral option parity for exllamav3
- Issue comment#414moshe44102026-02-22 18:59[BUG] Multiple missing dependencies in docker image
- Issue#414lee-b2026-02-16 01:41[BUG] Multiple missing dependencies in docker image
- Issue#414lee-b2026-02-16 01:41[BUG] Multiple missing dependencies in docker image
- Pull request#411lesj06102026-02-11 16:22
- Issue comment#405CNmnius2026-02-07 17:30[BUG] EXL3 reserve allocations ignored for device IDs > 0
- Issue comment#404thad0ctor2026-01-27 05:40[BUG] Exceptions on tool calls `tojson(ensure_ascii=False)`
- Issue#409baigarif1966-max2026-01-25 07:02[BUG]
- Issue#408baigarif1966-max2026-01-25 06:56[BUG]
- Issue comment#373baigarif1966-max2026-01-25 06:49add logprobs support for exl3
- Issue comment#373baigarif1966-max2026-01-25 06:48add logprobs support for exl3
- Issue comment#407Ph0rk0z2026-01-21 13:56Adaptive p sampling to exllamav3
- Issue comment#407turboderp2026-01-20 21:54Adaptive p sampling to exllamav3
- Issue comment#314chrisoutwright2026-01-12 01:14[BUG] Tool calling with Mistral doesn't work, even with correct prompt template.
- Issue comment#378mratsim2026-01-11 21:02feat: Add XML tool call processing system for GLM-4.5 and Qwen3-Coder models
- Issue comment#378torisetxd2026-01-11 15:13feat: Add XML tool call processing system for GLM-4.5 and Qwen3-Coder models
- Issue comment#295mratsim2026-01-09 09:51:sparkles: reasoning parser support
Totals cover only the window loaded into ClickHouse and count events, not GitHub's lifetime totals — 494 stars here means stars gained during the window, not the repo's star count.