Support · MiniCPM5-1B
Will it last?
Ownership levelPartialnone·limited·partial·substantial·fullAnalytical input C ยท 56.8/100
This page is a projection of the one entry record, the Transparency factor that Support covers. The full verdict is set by all four factors together, floor-weighted so the weakest caps the whole.
Which domain expands which factor
- AssessUse & modify + Transparency
- ImplementData control + Reliability
- UseReliability
- SupportTransparency
Common problems & fixes
The list is drawn from the model card and cookbooks (ev-card, ev-github) and is not exhaustive:
- Old stacks fail to load the model. The card sets floors of vLLM 0.21, SGLang 0.5.12 and transformers 5.6, so upgrade before debugging anything else.
- Tool calls come back as raw XML. Use SGLang with
--tool-call-parser minicpm5, or the repository's parser on other servers.
- Answers ramble or truncate. Set
enable_thinking explicitly and use the card's sampling for each mode (0.9 Think, 0.7 No Think, top_p 0.95). The publisher reports that RL and OPD cut max-token hits by 29 percentage points on its own set.
- A small device runs out of memory. Use the Q4_K_M GGUF or MLX 4-bit and cap the context (the cookbook example uses 8192).
- Older MiniCPM checkpoints are under the General Model License and not Apache-2.0 (
ev-legacy-gml).
Versions, changelog & cadence
The GitHub README log dates MiniCPM5-1B at 2026-05-19 and MiniCPM5-2B at 2026-09-07. Other sources say 25 May (ev-github, ev-aa), and the Hub repos were created 2026-05-21. The Hub repos show last-modified dates of 2026-11-09 to 2026-11-11. The cadence since 2024 is a new generation roughly every 8 to 12 months (Feb 2024, Sep 2024, Jun 2025, May 2026) plus point releases. The repository has no GitHub releases or tags, so the Hub commit hash is the changelog anchor; pin it.
GitHub issues with bug-report, bad-case and feature-request templates, GitHub Discussions, a Discord server, a Feishu or Lark group and a Chinese wiki at openbmb.cn/docs (ev-github). Commercial authorisation by email applies only to older generations.
Tracked known issues
No tracker exists for the model, so these come from the card and table (hence partial). There is no safety evaluation. The GPQA-Diamond, IFEval, Multi-IF, MultiChallenge and BFCLv4 rows are weaker than some baselines. No long-context result exists for the 1B, and the SFT dataset sits behind a Hub gate. The shared MiniCPM repository had 24 open issues on 2026-11-19, which are not triaged. Follow-ups are MiniCPM5-2B (separate release) and older MiniCPM generations (General Model License).
How this scores
The ownership factor this domain covers, drawn from the one entry record.
2
TransparencyDo you know what it is: weights, training, behaviour, and legible terms?
ModerateThe recipe and three pretraining datasets are public, but the SFT dataset card is gated and unread, the RL data is partly unreleased, there is no MiniCPM5 report or training pipeline, and the benchmark table is an unverified image.
How this scores (AOI sub-dimensions)
Provenance3/5how well we can trace and verify what went into the modelSafetensors only (no pickle), Hub file scan clean, per-file LFS sha256, official GGUF and MLX conversions, and the GitHub repository links these exact Hub repos; checklist about 5 of 8.
Governance3/5how accountable and well-documented the publisher isAn identifiable, accountable publisher group (Modelbest, THUNLP, Renmin University) with a visible cadence since 2024 (MiniCPM 2B Feb 2024, MiniCPM3 Sep 2024, MiniCPM4 Jun 2025, MiniCPM5-1B May 2026, MiniCPM5-2B Sep 2026), an active repository (last push 2026-11-09), GitHub issue templates and Discussions.
What this means for adoptionUse of these weights is fully owned: Apache-2.0, ungated, runnable on a phone or laptop, with Base and SFT checkpoints and public pretraining data to adapt from. Ownership is partial because the reliability factor is weak: OpenBMB publishes no safety evaluation, policy or guard model. The SFT data card is gated, and the only benchmark is the publisher's own table. The model suits local assistants and fine-tuning behind the deployer's own guardrails and safety tests. Ownership moves to substantial once a safety evaluation exists.
Sources
The same evidence records as the entry sheet. Read means the text was verified; unverified means it is known to exist but not yet read.
Licenceread2026-11-19
The LICENSE file in the OpenBMB/MiniCPM repository, which the MiniCPM5-1B model card links as the licence for "this repository and MiniCPM model weights", is the standard Apache License 2.0 text.
Licenceread2026-11-19
Hub card metadata for openbmb/MiniCPM5-1B, -SFT, -Base, -GGUF, -MLX and every MiniCPM5-2B checkpoint reads license apache-2.0 with gated false.
Licenceread2026-11-19
The legacy README for earlier MiniCPM generations says model weights "must strictly follow" a custom General Model License (GML), with commercial use needing authorisation from cpm@modelbest.cn by registration.
Model cardread2026-11-19
The MiniCPM5-1B model card describes a dense LlamaForCausalLM with 1,080,632,832 parameters, 24 layers, GQA with 16 query and 2 key-value heads, and a context length of 131,072.
Documentationread2026-11-19
The Hub repository holds one safetensors shard (2,161,290,912 bytes, LFS sha256 7ab8fd86...), no pickle and no LICENSE file.
Documentationread2026-11-19
The dataset cards for Ultra-FineWeb (about 1T English and 120B Chinese tokens), Ultra-FineWeb-L3 and UltraData-Math show that all three are public, ungated and Apache-2.0 tagged.
Documentationunverified2026-11-19
UltraData-SFT-2605 exists on the Hub (API metadata: gated auto, apache-2.0, 1,510 files, created 2026-05-21) but its card returns 401 "restricted ...
Technical_reportread2026-11-19
The card cites two arXiv papers.
Vendor announcementread2026-11-19
The publisher benchmark image on the model card (Thinking mode, three small baselines) reports an average of 42.57 vs Qwen3-0.6B 26.77, Qwen3.5-0.8B 25.14 and LFM2.5-1.2B 35.61.
Third-party analysisread2026-11-19
Artificial Analysis lists MiniCPM5-1B as an open-weights OpenBMB model, Apache 2.0, text only, 128k context, released 25 May 2026, with no publicly available API and no Intelligence Index score or other benchmark result on its model page.
Documentationread2026-11-19
The OpenBMB/MiniCPM repository holds a release log (MiniCPM5-1B 2026-05-19) and names the institutions (Modelbest, THUNLP, Renmin University GSAI).
Model cardread2026-11-19
The MiniCPM5-2B card (second model in the series, released 2026-09-07, Apache-2.0, ungated) lists a different and larger dataset set (adds UltraData-Code, UltraData-SFT-Agent-2609, UltraData-RL-2609, UltraX-Preview), so it is a separate release and is not scored in this entry.