Nanbeige releases Nanbeige4.3-33B open 33B agentic coding model
Original title: Nanbeige Releases Nanbeige4.3-33B, a 33B Open Agentic Coding Model
Apache 2.0 weights ship with a 524,288-token context and a base checkpoint; the lab says it leads 17 of 20 self-run benchmarks.
Nanbeige LLM Lab, inside recruitment platform BOSS Zhipin, published Nanbeige4.3-33B and Nanbeige4.3-33B-Base on Hugging Face under Apache 2.0. The dense 33B Llama-style model targets agentic coding and office cowork tasks, with a 524,288-token window and reasoning on by default. In the lab’s table it leads Qwen3.8-27B, Muse-Glimmer-30B and Gemma4-31B on 17 of 20 rows, including 80.9 on Terminal-Bench 2.1, but trails Qwen on SWE-Bench Pro. Several competitor figures conflict with those labs’ own model cards.
Unconfirmed: Benchmark scores are self-reported, and several competitor figures conflict with those labs’ published cards.
- BF16 weights need about 66 GB; recipes use 4-way tensor parallelism.
- Terminal-Bench 2.1 is 80.9 versus 73.0 for Qwen3.8-27B.
- SWE-Bench Pro is 60.3, behind Qwen3.8-27B at 61.7.
- The lab plans to distill agentic skills into Nanbeige4.3-3B.
- Chat controls cover thinking, preserved reasoning, and XML tool calls.
- Vendors
- Nanbeige LLM Lab
- Organizations
- BOSS Zhipin
From the original article
7 min read at marktechpost.com
Nanbeige LLM Lab released Nanbeige4. 3-33B, a 33B dense Apache 2. 0 model with 512K context.
By Asif Razzaq · marktechpost.com
A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.