Granite 4.2 30B vs InternVL3.5-GPT-OSS-20B-A4B-Preview

Granite 4.2 30B 87

Model card 17, Intended use (best for), limitations (ethical considerations and limitations), architecture, parameters (30B), context length (128K native) and 12 tested languages are on the card. Training data 20, IBM's disclosure file for the 30B lists the datasets used in pre-training and post-training as machine-readable metadata, and the card describes the three training stages. Last change 15, the last commit to the repository was on 4 September 2026, 67 days before this check. Responsiveness 20, Of the 4 discussions that aren't pull requests (28 August to 3 November 2026), 3 have a reply from an ibm-granite member. Version policy 5, Granite 4.1 30B stays in its own repository and 4.2 came out in new ones. File health 10, Config, template and card agree on the architecture, the 131,072-token context and the thinking switch. 17 + 20 + 15 + 20 + 5 + 10 = 87.

InternVL3.5-GPT-OSS-20B-A4B-Preview 28

Model card 8 of 20, the family card misses limitations, context, languages and a cut-off. Training data 5, described without a cut-off. Last change 0, the last commit to either repository is 8 September 2025 and to OpenGVLab's reference code 22 September 2025, over a year before 4 November 2026. Responsiveness 10, OpenGVLab answered 3 of 8 questions across the two repositories, none since September 2025. Version policy 5, a first release. File health 0, the template uses markers the tokenizer doesn't have and contradicts the twin's template, and the card names the wrong language model in its licence line.

Granite 4.2 30B 85

Chat template 20, chat_template.jinja in the repository, with roles, tools, tool responses, the thinking switch and low effort. Tool-calling format 25, the card has tool-calling sections for vLLM and SGLang, declares tools with OpenAI's function schema and names the parser to use, qwen3_coder, which both runtimes ship in the released versions read. Trained tool calling 20, the template loops over every entry in an assistant turn's tool_calls, so it renders several calls in one turn, and the card names tool calling and agentic workflows as what it's best for. Structured output 0, the card doesn't document JSON or JSON Schema output or how to ask for it. Reasoning control 10, the card documents thinking on by default, non-thinking and low-effort modes per request through the template (enable_thinking and low_effort), with the reasoning returned apart from the answer. Context 10, 131,072 tokens natively per the card and config.json. 20 + 25 + 20 + 0 + 10 + 10 = 85.

InternVL3.5-GPT-OSS-20B-A4B-Preview 45

Chat template 20, a Jinja file in the repository, though it doesn't match the tokenizer. Tool-calling format 10 of 25, only the twin's inherited harmony template mentions tools. Trained tool calling 0, no evidence it was trained to call tools. Structured output 0. Reasoning control 10, a system prompt turns thinking on and the card says the reasoning comes back in think tags. Context 5, trained at 32K tokens.

Granite 4.2 30B 90

Memory 25, IBM's own 4-bit MLX build granite-4.2-30b-q4-mlx is 16.47 GB, and 16.47 GB × 1.2 = 19.8 GB, which fits 24 GB but not 16 GB. Its Q4_0 GGUF is 16.58 GB and Ollama's library build granite4.2:30b 18 GB, the same band. Runtimes 25, Released versions run the granite architecture in llama.cpp (granite, b12242), vLLM (GraniteForCausalLM in the v0.33.0 registry), SGLang (granite.py at v0.5.23), Transformers (v5.22.2) and Ollama (its own granite4.2 library entry, 30b tag). Builds 15, IBM publishes FP8, MXFP4 and NVFP4 safetensors builds, a GGUF repository with Q4_0 to Q8_0 and BF16 files, and 4-bit, 8-bit and BF16 MLX builds, so quantised builds at two precisions or more. Terms 15, Apache 2.0, an OSI-approved licence, with no use policy or other terms added on the card. Download 10, the Hub API reports gated false for the BF16, FP8, MXFP4, NVFP4, GGUF and MLX repositories, so the files download with no account and nothing to accept. Hosted 0, Hugging Face maps one inference provider (DeepInfra, live), short of the two the line needs. 25 + 25 + 15 + 15 + 10 + 0 = 90.

InternVL3.5-GPT-OSS-20B-A4B-Preview 65

Memory 15 of 30, the only builds are BF16 at 42.5 GB by the repository's safetensors index, or 51 GB with a fifth added, which fits 64 GB. Runtimes 25, Transformers and vLLM load this repository, and SGLang, llama.cpp and Ollama (through llama.cpp b11963, which Ollama v0.40.14 pins) run the same weights in their GitHub format. Builds 0, OpenGVLab publishes no quantised build. Terms 15, Apache 2.0 with nothing added by the bases. Download 10, not gated. Hosted 0, no provider serves it.