All tasks / Tools
TabbyAPI (ExLlamaV3)
Open-source OpenAI-compatible server for the ExLlamaV2/V3 GPU inference libraries, focused on fast quantised models.
- theroyallab
- Local only
- Free, open source
Where it ranks
- AChat & assistantsRuntimes & gateways · Local · via GLM-5.3-Flash
- CCodingRuntimes & gateways · Local · via GLM-5.3-Flash
Tools without a benchmark of their own inherit the tier of the best model they offer.
Facts
- Vendor
- theroyallab
- Approach
- Runtimes & gateways
- Where it runs
- Runs only on your own hardware
- Interface
- API
- Skill level
- Advanced: command line, node graphs or self-hosting
- Price
- Free, open source
- Pricing notes
- Free and open source.
- Source code
- Open source (AGPL-3.0)
- GitHub stars
- 1.5k (19 Oct 2026)
- Last verified
- 19 Oct 2026
Price from github.com, 19 Oct 2026.
Models it gives you
- AGLM-5.3-FlashZ.ai · open weights
- AGemma 4 (31B)Google DeepMind · open weights
- AQwen3.8-27BAlibaba (Qwen) · open weights
- Bgpt-oss-120bOpenAI · open weights
- Cgpt-oss-20bOpenAI · open weights
- CQwen3-Coder-NextAlibaba (Qwen) · open weights
- –Devstral Small 2 (2512)Mistral AI · open weights