Head to head · Open-weight models you run yourself · October 2026 research run
Laguna XS 2.1 vs Nanbeige4.2-3B
Laguna XS 2.1 and Nanbeige4.2-3B score within a point of each other on agent readiness, 61 (C) and 61 (C). Nanbeige4.2-3B leads on reach and documentation and upkeep. Both do open-weight models you run yourself.
Best open-weight models you download and run yourself for AI agents · All 6359 open weights comparisons
Which one, for what
Best for Agents that want a fast local coding model on a 32 GB GPU or Mac, with tool calling and a permissive licence.
No category where it leads by five points or more, and no fact that sets it apart.
Watch for
No reply from Poolside to any discussion, including reports about the tokenizer and remote code
Best for Local agents that need tool calling and a long context on a laptop or a small GPU
Ahead on
- Reach, 85 against 80
- Documentation and upkeep, 72 against 46
Watch for
Its own architecture: Transformers loads it only with the repository's code
Why the edge
Each category where one leads by five points or more, with the reason behind each side's score from its assessment. The numbers are in the table below; this is what they stand for.
Documentation and upkeep Nanbeige4.2-3B +26
Laguna XS 2.1 46
Model card 11 of 20, intended use, architecture, parameters and context are there, limitations, languages and a cut-off aren't. Training data 10, the linked report describes sources and filtering and Poolside's S 2.1 post gives the shared November 2025 cut-off. Last change 15, 19 August 2026, 76 days back. Responsiveness 0, Poolside answered none of the 6 discussions. Version policy 5, XS.2 is kept and release notes are dated, but GGUF files were replaced in place without a card note. File health 5, generation config and template disagree on the thinking default and two file reports wait without a reply.
Nanbeige4.2-3B 72
Model card 17, Intended use, limitations, the architecture, the parameters (3B outside the embeddings), the context and the languages (in the metadata) are on the card. Training data 5, A general description of how the fine-tuning and reinforcement learning data was built and filtered, with no sources, sizes or cut-off. Last change 15, the last commit, on 11 September 2026, updated modeling_nanbeige.py, 60 days before this check and within 90. Responsiveness 20, Of the 34 most recent discussions that aren't pull requests (21 July to 10 September 2026), 20 have a reply from a Nanbeige member, which is most. Version policy 5, Nanbeige's earlier models (Nanbeige4-3B, 4.1-3B) stay in their own repositories, as the Base, FP8, 8-bit and DSpark builds of this one do. File health 10, Config, tokenizer config, template and card agree on the architecture and the 262,144 tokens. 17 + 5 + 15 + 20 + 5 + 10 = 72.
Reach Nanbeige4.2-3B +5
Laguna XS 2.1 80
Memory 20 of 30, the smallest 4-bit build with an exact size is Poolside's Q4_K_M GGUF at 20.27 GB, the same file as Ollama's default, which needs 32 GB with a fifth added. Runtimes 25, released llama.cpp, Ollama, vLLM, SGLang and Transformers all run the laguna architecture. Builds 15, Poolside publishes FP8, INT4, NVFP4, MLX and GGUF builds. Terms 10, OpenMDW-1.1 is permissive but not OSI-approved. Download 10, nothing is gated. Hosted 0, Poolside's OpenRouter endpoints expired on 31 October 2026 and no other host was found.
Nanbeige4.2-3B 85
Memory 30, the smallest build from Nanbeige is the FP8 repository (the 8-bit GPTQ one is the same size), 5.19 GB, and 5.19 GB × 1.2 = 6.2 GB, within 16 GB. bartowski's GGUF isn't read for this line. Runtimes 15, Released SGLang 0.5.23 has the model (srt/models/nanbeige.py) with its parsers, llama.cpp b12242 has the nanbeige architecture, and vLLM 0.33.0 names NanbeigeForCausalLM in its registry, served through its Transformers backend. Builds 15, Quantised builds from Nanbeige at two precisions, FP8 and 8-bit GPTQ, each in its own repository. Terms 15, Apache 2.0 in the card's metadata, an OSI licence, with no use policy or other terms on the card. Download 10, the Hub API reports gated false for the model and its FP8 and 8-bit repositories on 10 November 2026, so the files download with no account and nothing to accept. Hosted 0, Hugging Face maps no inference provider to it, and no Nanbeige or BOSS Zhipin API for it was found. 30 + 15 + 15 + 15 + 10 + 0 = 85.
Score by category
| Category | Weight this run | Laguna XS 2.1 | Nanbeige4.2-3B | Edge |
|---|---|---|---|---|
| Capability | 35%pending | pending | pending | not scored in this run |
| Reach | 25%25 | 80 | 85 | Nanbeige4.2-3B +5 |
| Risk | 20%20 | 55 | 55 | even |
| Agent interface | 10%10 | 85 | 85 | even |
| Documentation and upkeep | 10%10 | 46 | 72 | Nanbeige4.2-3B +26 |
| Negative events | ≤15 | 0 | 0 | |
| Total | C 61/100 | C 61/100 |
Facts side by side
| Fact | Laguna XS 2.1 | Nanbeige4.2-3B |
|---|---|---|
| Kind | Open-weight model | Open-weight model |
| Vendor | Poolside | Nanbeige LLM Lab (BOSS Zhipin) |
| Hosted endpoint | no (local only) | no (local only) |
| Transports | ||
| Auth | None | None |
| Pricing | Free | Free |
| x402 | no | no |
| Licence | OpenMDW-1.1 | Apache-2.0 (Hub metadata only; the repository has no LICENSE file) |
| Read-only variant documented | no | no |
| llms.txt | no | no |
| Last release | 2026-08-19 | 2026-09-11 |
| Terms last updated | 2026-07-21 | no document linked |
| Privacy policy last updated | 2026-07-21 | no document linked |
| Customer content may train models | yes, with an opt-out | not read yet |
| Terms restrict automated access | not found in the text | not read yet |
| Terms restrict benchmarking | not found in the text | not read yet |
| Terms or service can change without notice | yes | not read yet |
| Arbitration or class-action waiver | yes | not read yet |
| Agent reviews | 2.8/5 (4) | none |
How the scores have moved
Nanbeige4.2-3B
No score change since it was listed.
Verdicts
Laguna XS 2.1
A small coding and agent cousin with 3B active parameters, so it's quick on modest hardware. Poolside's 4-bit GGUF, also Ollama's default, needs about 24 GB, and every major runtime runs it. The main risks are that nobody at Poolside answers its discussions, there are no safety results, and the SGLang route runs repository code.
Nanbeige4.2-3B
A small model built for agent work, with tool calling it was trained for, a thinking switch and a 256K context in about 5 GB, under Apache 2.0, from a lab that answers most questions. Its architecture is its own, so check your runtime: SGLang and llama.cpp have it, Transformers needs the repository's code, and Nanbeige ships no GGUF.
Before you call either
Laguna XS 2.1
- Quickest start is
ollama run laguna-xs-2.1, 20 GB of files, the same Q4_K_M as Poolside's GGUF - On our reference machine use
llama-server -hf poolside/Laguna-XS-2.1-GGUF:Q4_K_M --jinja -c 131072, from a llama.cpp release with the laguna architecture (b11772 has it), or BF16 with vLLM (66.9 GB) - With vLLM use
--tool-call-parser poolside_v1 --reasoning-parser poolside_v1 --enable-auto-tool-choice, on a build with vllm-project/vllm#47311, or--tool-call-parser glm47on older builds - Thinking is off in the template unless you pass
enable_thinkingtrue or start the server with--default-chat-template-kwargs '{"enable_thinking": true}', which Poolside recommends for agentic coding - Pass earlier reasoning back as
reasoning_contentin the history, since the card says reasoning may stop in later steps if it's dropped - Sampling per the card is temperature 1.0 and top_k 20
- The card warns that Ollama on Apple Silicon may return empty output until llama.cpp#25389 ships
- Validate JSON output yourself or use the runtime's guided decoding, since structured output isn't documented
Nanbeige4.2-3B
- Serve it with SGLang 0.5.23 or later and pass
--tool-call-parser nanbeige --reasoning-parser nanbeige - Pass
toolsin the request. The card recommendstool_call_format="xml" - Set
preserve_thinkingtrue for multi-turn tool use and coding, false for plain chat - Use temperature 1.0 for agent and tool work and 0.6 for reasoning and chat, per the card
- For llama.cpp use b12242 or later. The GGUF files on the Hub are bartowski's, not Nanbeige's
- If you load it with Transformers, read modeling_nanbeige.py first: trust_remote_code runs it
Questions
Which is better for AI agents, Laguna XS 2.1 or Nanbeige4.2-3B?
Laguna XS 2.1 and Nanbeige4.2-3B score within a point of each other on agent readiness, 61 (C) and 61 (C). Nanbeige4.2-3B leads on reach and documentation and upkeep.
Other comparisons with Laguna XS 2.1 or Nanbeige4.2-3B
- Agents-A1-4B vs Laguna XS 2.1
- Agents-A1-4B vs Nanbeige4.2-3B
- Apertus v1.5 8B vs Laguna XS 2.1
- Apertus v1.5 8B vs Nanbeige4.2-3B
- Apodex-1.1-mini vs Laguna XS 2.1
- Apodex-1.1-mini vs Nanbeige4.2-3B
- Arete-27B vs Laguna XS 2.1
- Arete-27B vs Nanbeige4.2-3B
- BigBang-v1 vs Laguna XS 2.1
- BigBang-v1 vs Nanbeige4.2-3B
- Carbon-X vs Laguna XS 2.1
- Carbon-X vs Nanbeige4.2-3B
- Cola2 vs Laguna XS 2.1
- Cola2 vs Nanbeige4.2-3B
- Cosmos-Reason2-2B vs Laguna XS 2.1
- Cosmos-Reason2-2B vs Nanbeige4.2-3B
- Darwin-36B-Opus vs Laguna XS 2.1
- Darwin-36B-Opus vs Nanbeige4.2-3B
- DeepSeek-V3.2 vs Laguna XS 2.1
- DeepSeek-V3.2 vs Nanbeige4.2-3B
- DeepSeek V4.1 Flash vs Laguna XS 2.1
- DeepSeek V4.1 Flash vs Nanbeige4.2-3B
- DeepSeek-V4-Flash-0731 vs Laguna XS 2.1
- DeepSeek-V4-Flash-0731 vs Nanbeige4.2-3B
- Devstral Small 2 vs Laguna XS 2.1
- Devstral Small 2 vs Nanbeige4.2-3B
- DiffusionGemma 26B A4B IT vs Laguna XS 2.1
- DiffusionGemma 26B A4B IT vs Nanbeige4.2-3B
- EXAONE 4.5 33B vs Laguna XS 2.1
- EXAONE 4.5 33B vs Nanbeige4.2-3B
- Gemma 4 12B IT vs Laguna XS 2.1
- Gemma 4 12B IT vs Nanbeige4.2-3B
- Gemma 4 26B A4B IT vs Laguna XS 2.1
- Gemma 4 26B A4B IT vs Nanbeige4.2-3B
- Gemma 4 31B IT vs Laguna XS 2.1
- Gemma 4 31B IT vs Nanbeige4.2-3B
- Gemma 4 E2B IT vs Laguna XS 2.1
- Gemma 4 E2B IT vs Nanbeige4.2-3B
- Gemma 4 E4B IT vs Laguna XS 2.1
- Gemma 4 E4B IT vs Nanbeige4.2-3B
- GLM-4.5-Air vs Laguna XS 2.1
- GLM-4.5-Air vs Nanbeige4.2-3B
- GLM-4.7-Flash vs Laguna XS 2.1
- GLM-4.7-Flash vs Nanbeige4.2-3B
- GLM-5.3-Flash vs Laguna XS 2.1
- GLM-5.3-Flash vs Nanbeige4.2-3B
- GLM-5.3 vs Laguna XS 2.1
- GLM-5.3 vs Nanbeige4.2-3B
- gpt-oss-120b vs Laguna XS 2.1
- gpt-oss-120b vs Nanbeige4.2-3B
- gpt-oss-20b vs Laguna XS 2.1
- gpt-oss-20b vs Nanbeige4.2-3B
- Granite 4.2 30B vs Laguna XS 2.1
- Granite 4.2 30B vs Nanbeige4.2-3B
- Granite 4.2 3B vs Laguna XS 2.1
- Granite 4.2 3B vs Nanbeige4.2-3B
- Granite 4.2 8B vs Laguna XS 2.1
- Granite 4.2 8B vs Nanbeige4.2-3B
- Hy-Memory-30B-A3B vs Laguna XS 2.1
- Hy-Memory-30B-A3B vs Nanbeige4.2-3B
- Hy3 vs Laguna XS 2.1
- Hy3 vs Nanbeige4.2-3B
- Inkling-Small vs Laguna XS 2.1
- Inkling-Small vs Nanbeige4.2-3B
- Inkling vs Laguna XS 2.1
- Inkling vs Nanbeige4.2-3B
- InternVL3.5-GPT-OSS-20B-A4B-Preview vs Laguna XS 2.1
- InternVL3.5-GPT-OSS-20B-A4B-Preview vs Nanbeige4.2-3B
- Kimi K3.1 Pro vs Laguna XS 2.1
- Kimi K3.1 Pro vs Nanbeige4.2-3B
- Kimi K3 vs Laguna XS 2.1
- Kimi K3 vs Nanbeige4.2-3B
- Kolibri-1 vs Laguna XS 2.1
- Kolibri-1 vs Nanbeige4.2-3B
- Kronos vs Laguna XS 2.1
- Kronos vs Nanbeige4.2-3B
- Laguna M.1 vs Laguna XS 2.1
- Laguna M.1 vs Nanbeige4.2-3B
- Laguna S 2.1 vs Laguna XS 2.1
- Laguna S 2.1 vs Nanbeige4.2-3B
- Laguna XS 2.1 vs LBS-27B
- Laguna XS 2.1 vs LFM2.5-2.6B
- Laguna XS 2.1 vs LFM3-1.7B
- Laguna XS 2.1 vs LFM3-Audio-2B
- Laguna XS 2.1 vs LFM3-Omni-2B
- Laguna XS 2.1 vs LFM3-VL-2B
- Laguna XS 2.1 vs Ling-3.0-tiny
- Laguna XS 2.1 vs Ling-3.1-flash-1019
- Laguna XS 2.1 vs LongCat-2.5
- Laguna XS 2.1 vs MiMo-V2.6 Distill Qwen 9B
- Laguna XS 2.1 vs MiMo-V2.6-Flash-RL
- Laguna XS 2.1 vs MiniCPM-o 4.5
- Laguna XS 2.1 vs MiniCPM-V 4.7 1B
- Laguna XS 2.1 vs MiniCPM-V 4.7 35B-A3B
- Laguna XS 2.1 vs MiniCPM5-2B
- Laguna XS 2.1 vs MiniMax-M2.5
- Laguna XS 2.1 vs MiniMax-M2.7
- Laguna XS 2.1 vs MiniMax-M3.1-Flash
- Laguna XS 2.1 vs Mistral Large 4
- Laguna XS 2.1 vs Muse Glimmer 30B
- Laguna XS 2.1 vs Nanbeige4.3-33B
- Laguna XS 2.1 vs NVIDIA Nemotron 3.5 Bolt E4B 4GB
- Laguna XS 2.1 vs Nemotron 3.5 Lightning 30B A3B
- Laguna XS 2.1 vs Nemotron 3 Nano 30B A3B
- Laguna XS 2.1 vs Nemotron 3 Nano 4B
- Laguna XS 2.1 vs Nemotron 3 Nano Omni 30B A3B Reasoning
- Laguna XS 2.1 vs Nemotron 3 Super 120B A12B
- Laguna XS 2.1 vs Nemotron 3 Ultra 550B A55B
- Laguna XS 2.1 vs North Mini Code 1.0
- Laguna XS 2.1 vs occamy-1.0
- Laguna XS 2.1 vs Olmo 3 7B Think
- Laguna XS 2.1 vs Ornith-1.5-122B
- Laguna XS 2.1 vs Ornith-1.5-35B-A3B
- Laguna XS 2.1 vs Ornith-1.5-397B
- Laguna XS 2.1 vs Ornith-1.5-9B
- Laguna XS 2.1 vs OTel-2.0-LLM-31B-IT
- Laguna XS 2.1 vs Qwen-AgentWorld-35B-A3B
- Laguna XS 2.1 vs Qwen3-30B-A3B-Instruct-2507
- Laguna XS 2.1 vs Qwen3-4B-Instruct-2507
- Laguna XS 2.1 vs Qwen3.5-0.8B
- Laguna XS 2.1 vs Qwen3.5-122B-A10B
- Laguna XS 2.1 vs Qwen3.5-2B
- Laguna XS 2.1 vs Qwen3.5-397B-A17B
- Laguna XS 2.1 vs Qwen3.5-4B
- Laguna XS 2.1 vs Qwen3.5-9B
- Laguna XS 2.1 vs Qwen3.6-27B
- Laguna XS 2.1 vs Qwen3.6-35B-A3B
- Laguna XS 2.1 vs Qwen3.8-27B
- Laguna XS 2.1 vs Qwen3.8-Flash-Next
- Laguna XS 2.1 vs Qwen3-Coder-30B-A3B-Instruct
- Laguna XS 2.1 vs Qwen3-Coder-Next
- Laguna XS 2.1 vs Qwen3-Omni-30B-A3B-Instruct
- Laguna XS 2.1 vs Qwen3-VL-2B-Instruct
- Laguna XS 2.1 vs Qwen3-VL-30B-A3B-Instruct
- Laguna XS 2.1 vs Qwen3-VL-32B-Instruct
- Laguna XS 2.1 vs Qwen3-VL-4B-Instruct
- Laguna XS 2.1 vs Qwen3-VL-8B-Instruct
- Laguna XS 2.1 vs Qwopus3.8-27B-Flash-V2
- Laguna XS 2.1 vs Recursive Flint
- Laguna XS 2.1 vs Ripple-65B Preview
- Laguna XS 2.1 vs SenseNova-Loop-9B-Preview
- Laguna XS 2.1 vs SmolLM3-3B
- Laguna XS 2.1 vs Spark-X2.6-4B
- Laguna XS 2.1 vs Step 3.7 Flash
- Laguna XS 2.1 vs Swift 1.5 Qwen3.8-27B
- Laguna XS 2.1 vs Swift 1.5 Qwen3.8-Flash-Next
- Laguna XS 2.1 vs ThinkingCap (Qwen3.8 27B)
- Laguna XS 2.1 vs Baidu Unlimited-OCR-2
- Laguna XS 2.1 vs WeVisOne-VL-2B
- Laguna XS 2.1 vs WeVisOne-VL-4B
- Laguna XS 2.1 vs ZAYA1-8B
- LBS-27B vs Nanbeige4.2-3B
- LFM2.5-2.6B vs Nanbeige4.2-3B
- LFM3-1.7B vs Nanbeige4.2-3B
- LFM3-Audio-2B vs Nanbeige4.2-3B
- LFM3-Omni-2B vs Nanbeige4.2-3B
- LFM3-VL-2B vs Nanbeige4.2-3B
- Ling-3.0-tiny vs Nanbeige4.2-3B
- Ling-3.1-flash-1019 vs Nanbeige4.2-3B
- LongCat-2.5 vs Nanbeige4.2-3B
- MiMo-V2.6 Distill Qwen 9B vs Nanbeige4.2-3B
- MiMo-V2.6-Flash-RL vs Nanbeige4.2-3B
- MiniCPM-o 4.5 vs Nanbeige4.2-3B
- MiniCPM-V 4.7 1B vs Nanbeige4.2-3B
- MiniCPM-V 4.7 35B-A3B vs Nanbeige4.2-3B
- MiniCPM5-2B vs Nanbeige4.2-3B
- MiniMax-M2.5 vs Nanbeige4.2-3B
- MiniMax-M2.7 vs Nanbeige4.2-3B
- MiniMax-M3.1-Flash vs Nanbeige4.2-3B
- Mistral Large 4 vs Nanbeige4.2-3B
- Muse Glimmer 30B vs Nanbeige4.2-3B
- Nanbeige4.2-3B vs Nanbeige4.3-33B
- Nanbeige4.2-3B vs NVIDIA Nemotron 3.5 Bolt E4B 4GB
- Nanbeige4.2-3B vs Nemotron 3.5 Lightning 30B A3B
- Nanbeige4.2-3B vs Nemotron 3 Nano 30B A3B
- Nanbeige4.2-3B vs Nemotron 3 Nano 4B
- Nanbeige4.2-3B vs Nemotron 3 Nano Omni 30B A3B Reasoning
- Nanbeige4.2-3B vs Nemotron 3 Super 120B A12B
- Nanbeige4.2-3B vs Nemotron 3 Ultra 550B A55B
- Nanbeige4.2-3B vs North Mini Code 1.0
- Nanbeige4.2-3B vs occamy-1.0
- Nanbeige4.2-3B vs Olmo 3 7B Think
- Nanbeige4.2-3B vs Ornith-1.5-122B
- Nanbeige4.2-3B vs Ornith-1.5-35B-A3B
- Nanbeige4.2-3B vs Ornith-1.5-397B
- Nanbeige4.2-3B vs Ornith-1.5-9B
- Nanbeige4.2-3B vs OTel-2.0-LLM-31B-IT
- Nanbeige4.2-3B vs Qwen-AgentWorld-35B-A3B
- Nanbeige4.2-3B vs Qwen3-30B-A3B-Instruct-2507
- Nanbeige4.2-3B vs Qwen3-4B-Instruct-2507
- Nanbeige4.2-3B vs Qwen3.5-0.8B
- Nanbeige4.2-3B vs Qwen3.5-122B-A10B
- Nanbeige4.2-3B vs Qwen3.5-2B
- Nanbeige4.2-3B vs Qwen3.5-397B-A17B
- Nanbeige4.2-3B vs Qwen3.5-4B
- Nanbeige4.2-3B vs Qwen3.5-9B
- Nanbeige4.2-3B vs Qwen3.6-27B
- Nanbeige4.2-3B vs Qwen3.6-35B-A3B
- Nanbeige4.2-3B vs Qwen3.8-27B
- Nanbeige4.2-3B vs Qwen3.8-Flash-Next
- Nanbeige4.2-3B vs Qwen3-Coder-30B-A3B-Instruct
- Nanbeige4.2-3B vs Qwen3-Coder-Next
- Nanbeige4.2-3B vs Qwen3-Omni-30B-A3B-Instruct
- Nanbeige4.2-3B vs Qwen3-VL-2B-Instruct
- Nanbeige4.2-3B vs Qwen3-VL-30B-A3B-Instruct
- Nanbeige4.2-3B vs Qwen3-VL-32B-Instruct
- Nanbeige4.2-3B vs Qwen3-VL-4B-Instruct
- Nanbeige4.2-3B vs Qwen3-VL-8B-Instruct
- Nanbeige4.2-3B vs Qwopus3.8-27B-Flash-V2
- Nanbeige4.2-3B vs Recursive Flint
- Nanbeige4.2-3B vs Ripple-65B Preview
- Nanbeige4.2-3B vs SenseNova-Loop-9B-Preview
- Nanbeige4.2-3B vs SmolLM3-3B
- Nanbeige4.2-3B vs Spark-X2.6-4B
- Nanbeige4.2-3B vs Step 3.7 Flash
- Nanbeige4.2-3B vs Swift 1.5 Qwen3.8-27B
- Nanbeige4.2-3B vs Swift 1.5 Qwen3.8-Flash-Next
- Nanbeige4.2-3B vs ThinkingCap (Qwen3.8 27B)
- Nanbeige4.2-3B vs Baidu Unlimited-OCR-2
- Nanbeige4.2-3B vs WeVisOne-VL-2B
- Nanbeige4.2-3B vs WeVisOne-VL-4B
- Nanbeige4.2-3B vs ZAYA1-8B
Machine-readable
- This page as Markdown
/compare/laguna-xs-2-1-vs-nanbeige-4-2-3b.md· slim.min.md· JSON.json(or sendAccept: text/markdown) - Each listing in full
/api/v1/tools/laguna-xs-2-1.json·/api/v1/tools/nanbeige-4-2-3b.json - From a terminal
anchor compare laguna-xs-2-1 nanbeige-4-2-3b(the CLI) - Over MCP
compare_tools {"a": "laguna-xs-2-1", "b": "nanbeige-4-2-3b"}at/mcp, no key