Nanbeige releases Nanbeige4.3-33B open model with 512K context for agentic coding
Nanbeige introduces Nanbeige4.3-33B, a ~30B-parameter open model released under Apache-2.0 and positioned as a flagship teacher model for agentic capabilities. The model card reports benchmark results against Qwen3.8-27B, Muse-Glimmer-30B, and Gemma4-31B on agentic coding and cowork tasks, and states a context length of up to 524,288 tokens (512K). It supports Transformers, SGLang, and vLLM deployment with tool-use and reasoning chat templates.
Open original source huggingface.coOne primary source; other links are context, not independent corroboration.
- Primary sourceX discovery · automated brief
- Discovery linkX
Grok Bot submitted the link. AINewsIQ fetched the primary source and prepared a source-based brief automatically. An editor has not reviewed the story; social attention and the brief do not independently verify its claims.
- Original repository+40
- Feed text or abstract+15
- No editor link review0
- Original non-discussion link+5
This measures the stored source trail and editorial steps. It does not certify accuracy, peer review, or independent corroboration. Popularity adds no points.
View full scoring →Read the scoring rubric →What the source says
Nanbeige introduces Nanbeige4.3-33B, a ~30B-parameter open model released under Apache-2.0 and positioned as a flagship teacher model for agentic capabilities. The model card reports benchmark results against Qwen3.8-27B, Muse-Glimmer-30B, and Gemma4-31B on agentic coding and cowork tasks, and states a context length of up to 524,288 tokens (512K). It supports Transformers, SGLang, and vLLM deployment with tool-use and reasoning chat templates.
A READER'S EVIDENCE CHECKHow to read this signalSix questions about claims, sources, risks, and open checks
This is a signal-level guide until claim-level evidence is available. “Not recorded” means the evidence has not been assessed here; it does not mean a result is false. How we check AI news →
- 01Who is making each claim?
The claimant and whether each claim was measured or asserted are not recorded in this public brief.
- 02What was actually tested?
Not recorded per claim. Check the source for sample size, tasks, model and benchmark versions, and reproducibility.
- 03Where are the sources?
Original link: huggingface.co. Brief basis: Hugging Face model card. Primary page date and quote locations are not established in this panel.
- 04What could mislead?
Benchmark results are self-reported by the model authors, with some evaluations using custom judges and harnesses; source claims not independently checked.
- 05What is at risk if wrong?
Not assessed per claim. Consider safety, privacy, cost, and adoption decisions before relying on the result.
- 06What would change the view?
No claim-specific check is recorded. An inspectable primary method or independent replication could change the assessment.
Key findings from the stored brief
- Model card reports leading scores on most listed agentic coding and cowork benchmarks versus three ~30B-class peers.
- Reports 80.9 on Terminal-Bench 2.1 versus 73.0 for Qwen3.8-27B, per the card's table.
- Supports a context length of up to 524,288 tokens (512K).
- Released under Apache-2.0 with Transformers, SGLang, and vLLM serving instructions.
Benchmark results are self-reported by the model authors, with some evaluations using custom judges and harnesses; source claims not independently checked.
This is a source-based abstract, not independent validation of the source's claims.
Evidence & sources
These links show where the signal came from and what the source page linked to. A linked page is context, not independent corroboration.
These are Grok Bot discovery notes. AINewsIQ prepared a brief from available source material. The bot's rationale, linked posts, authors, and stances have not been independently verified.
The Hugging Face card (repo created Oct 24, announced Oct 25) describes a 33B model with 512K context, built as the teacher for a coming 3B model. Nanbeige reports 80.9 on Terminal-Bench 2.1 and 23.8 on Terminal-Bench 4.0. Qwen3.8-27B still leads on SWE-Bench Pro and GDPval in its own table. All scores are self-reported.
Benchmark results are self-reported by the model authors, with some evaluations using custom judges and harnesses; source claims not independently checked.
- Bot tagged: questions Points out the widely quoted 80.9 is on the older Terminal-Bench 2.1; Nanbeige reports 23.8 on Terminal-Bench 4.0. Inspect post ↗
- Bot tagged: questions Notes the published scores are Nanbeige's own and independent evals are still needed. Inspect post ↗
Engagement counts, when shown, are observed snapshots. They do not establish accuracy or corroboration.
- x post ↗ · 1,187 likes, 155 reposts, 107 replies, 116 quotes · observed 10/25/2026, 4:02:41 PM
- x post ↗ · 42 likes, 2 reposts, 5 replies, 0 quotes · observed 10/25/2026, 4:02:41 PM
- x post ↗ · 1 likes, 0 reposts, 1 replies, 0 quotes · observed 10/25/2026, 4:02:41 PM
- reddit post ↗
Related archive coverage
These relationships record a citation or topic overlap, not confirmation of the finding.
Grok Bot submitted the link. AINewsIQ fetched the primary source and prepared a source-based brief automatically. An editor has not reviewed the story; social attention and the brief do not independently verify its claims.
Timeline & updates
- Signal dated by source or discovery feed
This timestamp places the signal in the radar. It may be the discovery post date rather than the primary page's first publication date.
- Shared brief stored
AINewsIQ prepared one abstract from Hugging Face model card.
A publication revision history is not stored for this signal. New engagement snapshots do not mean the source itself changed.
Why this may matter
An openly licensed model with long context and reported agentic coding results could matter for developers building coding and office-automation agents. The model also serves as a teacher for planned smaller distilled models.
Limits and open questions
Benchmark results are self-reported by the model authors, with some evaluations using custom judges and harnesses; source claims not independently checked.
Organization-specific financial, security, and operational effects require the organization's own context and review.
Discussion
Comments are reviewed before publication. Discuss the evidence, not the person.
No published comments yet.
Reader discussion opens with accounts. You can read the original source and published comments now.