Qwen3.8-27B, Muse Glimmer-30B, and K2-Horizon on 16, 24, and 32 GB

What fits

Run your own context length through the VRAM calculator. The ranges above are weights, not weights plus a 32K chat. The 15 October pair is written up on the Nemotron 3.5 Bolt page and the 64 GB to 128 GB guide.

Qwen3.8-27B

Qwen/Qwen3.8-27B is a 27B dense model. The Hub license field is Apache-2.0. The card describes a vision-language model with a 262,144-token native context. On 3 October 2026 the short-context formula was 27 × 0.5 × 1.2 + 2 = 18.2 GB. Unsloth’s GGUF tree the same day showed UD-Q4_K_S at 15.4 GB, Q4_0 at 16.1 GB, UD-Q4_K_M at 16.5 GB, and UD-Q4_K_XL at 17.6 GB. Those files are the weights. The formula is larger because of the 1.2 factor and 2 GB. A 16 GB card does not have room for KV cache on top of a 17.6 GB file. 24 GB is the discrete tier. 32 GB is headroom for a longer chat, not a different model. Ollama’s library listed qwen3.8:27b at 18 GB.

The older Qwen3 hardware guide still covers the 8B/14B/32B generation. qwen3:32b is not this model. The tag is qwen3.8:27b. Gemma 4’s comparable local MoE is on the what-can-I-run guide: 17.1 GB at Q4 when you count all 25.2B parameters. Kolibri-1, also dated 3 October, is about 78 GB FP8 and is a datacenter serve, not a row in the table above.

Muse Glimmer-30B

meta-models/Muse-Glimmer-30B is a ~29.6B dense model with a perception encoder. The model card says Apache-2.0 and an August 2026 release. Their own fit table is the one to trust: a K-Quant around 17 GB aimed at 24 GB, a dynamic quant aimed at 32 GB, and full precision aimed at 64 GB. They also say the 4-bit language-model weights come in under 20 GB, with KV cache and the vision encoder on top. A 16 GB gaming card is below the target they published.

Meta's card reports generation speed for the 17 GB quant plus their DFlash drafter: 74.9 tok/s baseline and 233.4 tok/s with speculation on an RTX 5090, 23.7 / 37.8 on an M4 Max, measured by them with batch size 1 and greedy decoding. Those are Meta's numbers, not ours. Speculation changes the picture; do not compare that 233 figure to a plain Q4 llama.cpp run of a different model.

K2-Horizon

IFM/K2-Horizon-7B is the dense 7B-class model (the Hub sidebar lists about 9B parameters, which usually includes embeddings). The card calls the release fully open — weights, data, and code — and dates the artifact index 1 October 2026. Read the LICENSE file before you ship a product on it. Q4 is about 4–5 GB, so an 8 GB GPU is enough for the weights. It has a 512K context on the card; you will not get 512K on 8 GB, because the KV cache will not fit. K2-3.7B is the smaller sibling if 8 GB feels tight once you turn context up.

K2-Horizon-MoVA-36B-A4B is the 36B / ~4B-active MoE from the same family (3 September 2026 in the research pass, Apache-2.0 listed there — confirm the file). Q4 of the full expert set is about 20–24 GB. Speed can resemble a small model. Memory cannot. A 16 GB card does not hold it. 24 GB is the minimum discrete tier, and it will be tight with context.

K2-32B, if you grab that checkpoint, behaves like other dense ~32B models: Q4 around 20 GB of weights, so 24 GB again, 32 GB if you want Q5/Q6 or a long chat.

The 24 GB card, and what it costs now

For this set, 24 GB is the useful discrete step: Qwen3.8-27B Q4, Muse's 17 GB quant, and MoVA Q4. A used RTX 3090 or a 4090 does that job. An RTX 5090 adds 32 GB and a lot of bandwidth. Extra VRAM does not make a 7B model faster once it already fits, and 32 GB still does not fit Qwen3.8-Flash-Next. The 5090 vs 4090 guide has the bandwidth comparison; the price is why upgrading "for tok/s" on an 8B makes little sense.

Live price only. The May catalog number is not repeated here.

On a Mac, the same 24–36 GB band was in stock on Amazon on 1 October 2026: Mac mini M5 Pro 24 GB and M4 Pro 24 GB for Qwen3.8-27B Q4 (tight) and K2-7B, and Mac Studio M5 Max 36 GB for Qwen3.8-27B plus K2-32B / 36B-A4B. DeepSeek-V4-Flash at an aggressive quant is the EVO-X2 (128 GB at ~256 GB/s) or an Apple Ultra, not those minis and not a 64 GB DDR5 box at ~90 GB/s. The Beelink GTR9 Pro matches the EVO-X2’s bandwidth and adds a NIC caveat; the ranking is in the Strix Halo guide. Mac prices and the out-of-stock 48 GB / 64 GB listings are in the Apple Silicon guide.

Also on the list, with caveats

• DeepSeek-V4-Flash-0731. A Flash build can be the practical download. It is still not "one 4090." See the DeepSeek guide.
• DeepSeek-V4.1-Flash-1017. On 17 October 2026 the safetensors total was 763,205,315,794 parameters and the files summed to 510.3 GB. Q4 on 763.2B is 459.9 GB. That file does not fit 128 GB, 256 GB, or 512 GB after an OS reserve. It is not the older Flash-0731 checkpoint. Hugging Face card. See the DeepSeek guide.
• Qwen3.8-Flash-Next. Card: 125B with about 6B active, plus a 51B n-gram embedding and a 4B MTP head. On 9 October 2026 Unsloth UD-IQ4_XS was 93.68 GB and UD-Q4_K_XL was 111.33 GB. The formula on 180B is 110 GB. A 32 GB GPU does not fit it. A 128 GB EVO-X2 fits the IQ4 file and is tight on Q4. The LICENSE file is the Qwen Community License, not Apache-2.0. Qwen3.8-27B is the Apache-2.0 dense model. Qwen Cloud’s Flash API is the hosted product. File table: Flash-Next hardware. Hugging Face card.
• GLM-5.3-Flash. Useful only if the weights you downloaded match the card you read. This guide does not assign it a VRAM number from a title.
• MiniMax-M3. Weights were up around 11 June 2026. Do not treat the license as commercial-friendly until you open the LICENSE file on that repo.

15 October 2026: Bolt and Muse Spark Open

Nemotron 3.5 Bolt is 74B total and 9B active. NVIDIA’s card says the shipped local checkpoint stores experts at about 2 bits so it fits a 24 GB budget at batch size 1 and 128K context. Our Q4 formula on the total is 46.4 GB, which does not fit 24 GB or 32 GB. On 17 October 2026 Ollama listed nemotron-3.5-bolt:latest at 25GB. That download is the edge of a 24 GB card and may spill. Do not load a Q4. The parent Super VL model is 120B total. The Q4 formula is 74 GB. Ollama’s 120b-a12b-q4_K_M tag was 90GB the same day: 128 GB fits, and 96 GB is the edge after macOS.

Muse Spark Open is about 816B total and about 42B active. It is not Muse Glimmer. Q4 is 491.6 GB. The Hugging Face llama.cpp post says BF16 is about 1.5 TB, and that a 2-bit quant used 304 GB and 18 tokens per second on a 512 GB M3 Ultra. That speed is theirs. Unsloth says start with dynamic 2-bit on 256 GB. A 32 GB GPU and a 128 GB mini PC do not run it. Weights that are still missing, including Bolt E4B (NVIDIA said 23 October) and Mistral Large 4, stay off this table.

17 October 2026: Beam, Granite Switch, and Dory

Reflection Beam is 501.3B total and 23.4B active, Apache-2.0, text, 1M context. Q4 is 302.8 GB, so 64 GB, 96 GB, and 128 GB all miss it. Unsloth UD-Q4_K_XL summed to 306.5 GB and misses 256 GB. UD-IQ1_S summed to 103.1 GB: that 1-bit file can sit on 128 GB and does not fit 96 GB. ollama run beam is beam:cloud, not a local weight. Size another quant in the VRAM calculator.

Granite Switch 4.1 8B is 8.78B. Q4 is 7.3 GB and fits 8 GB. Q8 is 12.5 GB and needs 16 GB. Granite Switch 4.1 30B is 29.99B. Q4 is 20.0 GB, so it misses 16 GB and fits 24 GB. Q8 is 38.0 GB and misses 32 GB. There is no ollama run granite-switch tag. ollama run granite4.1:8b is the older base, not this Switch bundle.

Dory 1.0 4B is 3.98B, OpenMDW 1.1, 256K context. Q4 is 4.4 GB and Q8 is 6.8 GB, so both fit 8 GB. The published Q5_K_M GGUF was 3,673,288,864 bytes. NVIDIA’s card says to use a llama.cpp fork, not a stock Ollama tag. The 24 GB side of Bolt is on the Bolt hardware page, and the 64–128 GB miss for Beam is on the 64 GB to 128 GB guide.

Announced, not a recommendation

Nex-N2.5-Pro, DeepSeek-V4.1-Pro, and MiMo-V2.6 showed up as announcements or incomplete drops. This guide does not say which GPU "runs" them. When a repo has files, a license, and a size, it can go in the calculator. Until then it is a press post.

Questions