# Cola2 vs Nanbeige4.3-33B (slim)

> Nanbeige4.3-33B scores 60.8 (C) to Cola2's 33.9 (F) for open-weight models you run yourself. Prices, MCP, x402, uptime and agent notes side by side.

- Full: https://www.anchorterminal.com/compare/cola2-vs-nanbeige-4-3-33b.md (~8,650 tokens) · this version ~630 tokens · JSON https://www.anchorterminal.com/compare/cola2-vs-nanbeige-4-3-33b.json · canonical https://www.anchorterminal.com/compare/cola2-vs-nanbeige-4-3-33b
- Index: https://www.anchorterminal.com/llms.txt · API: https://www.anchorterminal.com/api/v1/index.json · MCP: https://www.anchorterminal.com/mcp · Updated: 2026-11-10
- Licence: CC BY 4.0, credit Anchor Terminal and link this page (https://www.anchorterminal.com/terms/#data)

Nanbeige4.3-33B scores 60.8 (C) on agent readiness against Cola2's 33.9 (F), and leads in 3 of 4 scored categories. Both do open-weight models you run yourself.

- Cola2: F 33.9, rank #160 of 160 open-weight models · https://www.anchorterminal.com/tools/cola2.min.md
- Nanbeige4.3-33B: C 60.8, rank #86 of 160 open-weight models · https://www.anchorterminal.com/tools/nanbeige-4-3-33b.min.md
- Cola2, good for Researchers working on latent or diffusion language models who want a checkpoint to study
- Nanbeige4.3-33B, good for Agents with a 32 GB GPU (INT4) or more, or the 128 GB reference machine at BF16, that want an Apache-tagged 33B model for long coding and tool loops. Ahead on Reach, 80 against 25; Risk, 60 against 40; Agent interface, 85 against 20.

| Category | Cola2 | Nanbeige4.3-33B |
| --- | --- | --- |
| Capability (35%) | pending | pending |
| Reach (25%) | 25 | 80 |
| Risk (20%) | 40 | 60 |
| Agent interface (10%) | 20 | 85 |
| Documentation and upkeep (10%) | 58 | 56 |

| Fact (where they differ) | Cola2 | Nanbeige4.3-33B |
| --- | --- | --- |
| Vendor | ByteDance (Seed) | Nanbeige LLM Lab (BOSS Zhipin) |
| Licence | None stated for the weights (the code is Apache 2.0) | Apache-2.0 (Hub metadata only; the repository has no LICENSE file) |
| Last release | 2026-11-06 | 2026-10-31 |

Why the edge:
- Agent interface, Nanbeige4.3-33B +65: Cola2 20 (Chat template 20, A chat template in tokenizer_config.json (4,116 characters), which came with the Qwen3 tokenizer.); Nanbeige4.3-33B 85 (The chat template earns 20, Jinja in tokenizer_config.json.)
- Reach, Nanbeige4.3-33B +55: Cola2 25 (Memory 15, the only build is the F32 safetensors set, 35.07 GB, and 35.07 GB × 1.2 = 42.1 GB, within 64 GB.); Nanbeige4.3-33B 80 (Memory earns 20, since Nanbeige's INT4 build is 22.57 GB, about 27 GB with a fifth added, which fits 32 GB.)
- Risk, Nanbeige4.3-33B +20: Cola2 40 (File safety 15, safetensors only, no .bin, .pt or .pth files (25).); Nanbeige4.3-33B 60 (File safety earns 25, safetensors only and no repository code, so the card's trust_remote_code has nothing to run.)