Granite Speech 5.0 2B vs Nemotron 3.5 ASR Streaming 0.6B

Granite Speech 5.0 2B 95

Memory 30, the smallest build listed that keeps 4 bits is IBM's Q4_K_M GGUF, 1.14 GB with the 1.16 GB audio projector, and 2.30 GB × 1.2 = 2.8 GB, within 16 GB. IBM recommends Q8_0, 3.12 GB with the projector. Runtimes 20, Released Transformers 5.22.2 and vLLM 0.33.0 have the granite_speech model, llama.cpp b12242 has its converter and audio front end, and Ollama carries it as ibm/granite5-speech, which the card says was tested with v0.40.15. Builds 15, IBM publishes GGUF builds at fourteen quantised precisions, from Q2_K to Q8_0, with the audio projector. Terms 15, Apache 2.0 in the metadata and on the card, with nothing added. Download 10, the Hub API reports gated false for both repositories on 10 November 2026, so the files download with no account and nothing to accept. Hosted 5, IBM runs a Hugging Face Inference Endpoint that serves this model through an OpenAI-compatible API, with the address and code on the card. 30 + 20 + 15 + 15 + 10 + 5 = 95.

Nemotron 3.5 ASR Streaming 0.6B 70

Memory earns the full 30 because NVIDIA's own Q8_0 GGUF is 0.74 GB and even the F32 safetensors need only about 3 GB. Runtimes get 5, for Transformers from v5.13.0; vLLM points to a third party's plugin, SGLang and llama.cpp have only a Parakeet encoder for multimodal models, and Ollama and LM Studio have nothing. NVIDIA's NeMo and NeMo-Speech.cpp run it but aren't among the six. Builds get 10 for one quantised build, the Q8_0 GGUF. Terms get 10 because OpenMDW-1.1 is permissive with nothing added but isn't OSI-approved. Download gets 10 for an ungated repository and hosted too gets 5 for fal, Together and DeepInfra.

Granite Speech 5.0 2B 92

Model card 17, Intended use, limitations, the architecture, the size, the 16K positions and two-minute segments, and the languages are on the card. Training data 20, the data is listed with sources and sizes: a 39-row table giving each corpus, its task, its hours and a link to where it comes from, with the synthetic sets described. Last change 20, the last commit, on 6 November 2026, updated the card, 4 days before this check. Responsiveness 20, Both discussions that aren't pull requests (21 and 28 October 2026) have a reply from an ibm-granite member. Version policy 5, granite-speech-4.1-2b stays in its own repository and 5.0 came out in a new one, which names its predecessor and gives a release date. File health 10, Config, template and card agree on the architecture and the 16K positions. 17 + 20 + 20 + 20 + 5 + 10 = 92.

Nemotron 3.5 ASR Streaming 0.6B 69

The card with its subcards covers intended use, limitations, architecture, parameters and languages, and misses a context length and a cut-off (14). Training data names the sources and a size range without a cut-off (5). The files last changed on 10 September 2026, and NVIDIA's NeMo-Speech.cpp changed this model's cache-aware RNNT streaming path on 24 October 2026, released in v0.3.0 on 30 October (20). NVIDIA answered 18 of the 25 questions and reports (20). It's a first release, with the GGUF replaced in place through a titled pull request (5). An unanswered report of a tokenizer fault for Greek costs file health half (5).

Granite Speech 5.0 2B 70

File safety 25, safetensors and GGUF only, no .bin, .pt or .pth files and no Python in either repository. Provenance 15, the files sit in the ibm-granite organisation, which IBM's Granite docs link to (10). Safety evaluation 0, no safety results on the card. Agent guidance 10, the card says what the model does with a prompt it doesn't know: it ignores an unfamiliar or malformed prompt and transcribes, which it says limits how audio can steer the system. Licence risk 15, Apache 2.0 with no use policy, no threshold and no right to change the terms. Security programme 5, www.ibm.com/.well-known/security.txt names IBM PSIRT, read for the Granite 4.2 3B listing on 4 November 2026 (5). 25 + 15 + 0 + 10 + 15 + 5 = 70.

Nemotron 3.5 ASR Streaming 0.6B 60

File safety gets 15, because the NeMo checkpoint beside the safetensors is a pickle that one documented path loads; no repository code runs. Provenance gets 15 for files in NVIDIA's verified organisation with hashes for every file. Licence risk keeps 15, since OpenMDW-1.1 has none of the deductions and the one-day metadata slip after release was a correction. The security programme gets 10 for the card's vulnerability route and NVIDIA's PSIRT bulletins. Agent guidance gets 5 for general accuracy cautions, and safety evaluation 0, since no results are published.

Granite Speech 5.0 2B 55

Chat template 20, chat_template.jinja in the repository (193 characters). Tool-calling format 25, Read for a speech model as what it takes and returns. Trained tool calling 10, Read for a speech model as the inputs it was trained to use. Structured output 0, Read for a speech model as typed outputs. Reasoning control 0, A speech model with no reasoning. Context 0, the language model is extended to 16K positions, per the card and config.json (16,384), below the 32K band, and the card says to send audio in segments of about two minutes. 20 + 25 + 10 + 0 + 0 + 0 = 55.

Nemotron 3.5 ASR Streaming 0.6B 50

Read for a speech model, the tool-calling, trained tool calling and structured-output lines stand for what it takes and returns. NVIDIA documents the audio, the language prompt and the chunk settings, and the language-tagged output with how to strip the tag in Transformers and NeMo (25). It was trained for one stream with one language prompt at a time (10). Timestamps and confidence scores are claimed in a subcard with no way shown to ask for them (5). The chat template line gets 10 because the language prompt is given only as processor or script arguments. Reasoning control and context score 0, with no reasoning and no stated context.