RunsOnMy ModelsGPUs & MacsBy VRAMAPI pricesVRAM calculatorNew

Can I run granite 4.2 8b locally?

granite 4.2 8b by ibm-granite has 8.8B parameters. At 4-bit with an 8K context it needs about 6.8 GB of GPU memory. It fits on 65 of 65 GPUs and Macs we track. See all models that fit in 8 GB of VRAM.

6.8 GBQ4_K_M · 8K context
RTX 4060smallest consumer GPU that fits
M1 16GBsmallest Mac that fits
128Kmax context
GQAattention type

Memory needed by quant and context length

GB of GPU memory = weights + KV cache (FP16) + runtime overhead. Weights use real file sizes from ibm-granite/granite-4.2-8b-GGUF where available (marked •), otherwise parameter count × bits per weight. How we calculate.

QuantQuality4K ctx8K ctx32K ctx128K ctx
2-bit (Q2_K) •rough4.4 GB5.0 GB8.8 GB23.8 GB
3-bit (Q3_K_M) •usable5.3 GB5.9 GB9.6 GB24.6 GB
4-bit (Q4_K_M) •good, most popular6.2 GB6.8 GB10.6 GB25.6 GB
5-bit (Q5_K_M) •very good7.0 GB7.7 GB11.4 GB26.4 GB
6-bit (Q6_K) •near lossless7.9 GB8.6 GB12.3 GB27.3 GB
8-bit (Q8_0) •effectively lossless9.9 GB10.5 GB14.3 GB29.3 GB
16-bit (F16/BF16) •original weights17.6 GB18.2 GB22.0 GB37.0 GB

Which GPUs and Macs run granite 4.2 8b?

Best quality quant that fits at 8K context. "Tight" = within 10% of usable memory. Speed is a ceiling from memory bandwidth (real speed is typically 50–80% of it).

HardwareMemoryFits?Best quantNeedsMax tok/s (est.)
RTX 3060 12GB12 GBFitsQ6_K8.6 GB50details
RTX 4060 Ti 16GB16 GBFitsQ8_010.5 GB31details
RTX 309024 GBFitsF1618.2 GB53details
RTX 409024 GBFitsF1618.2 GB57details
RTX 509032 GBFitsF1618.2 GB102details
RTX 5070 Ti16 GBFitsQ8_010.5 GB96details
RX 7900 XTX24 GBFitsF1618.2 GB55details
Mac M4 Pro 48GB48 GBFitsF1618.2 GB16details
Mac M4 Max 128GB128 GBFitsF1618.2 GB31details
Mac M3 Ultra 512GB512 GBFitsF1618.2 GB47details
DGX Spark128 GBFitsF1618.2 GB16details
RTX 407012 GBFitsQ6_K8.6 GB70details
RTX 5060 Ti 16GB16 GBFitsQ8_010.5 GB48details
RTX 40608 GBFitsQ4_K_M6.8 GB51details
Arc B58012 GBFitsQ6_K8.6 GB63details
RTX 3060 Ti8 GBFitsQ4_K_M6.8 GB84details
RTX 30708 GBFitsQ4_K_M6.8 GB84details
RTX 3080 10GB10 GBFitsQ5_K_M7.7 GB122details
RTX 3080 12GB12 GBFitsQ6_K8.6 GB126details
RTX 3090 Ti24 GBFitsF1618.2 GB57details
RTX 4060 Ti 8GB8 GBFitsQ4_K_M6.8 GB54details
RTX 4070 Super12 GBFitsQ6_K8.6 GB70details
RTX 4070 Ti12 GBFitsQ6_K8.6 GB70details
RTX 4070 Ti Super16 GBFitsQ8_010.5 GB72details
RTX 408016 GBFitsQ8_010.5 GB77details
RTX 4080 Super16 GBFitsQ8_010.5 GB79details
RTX 50608 GBFitsQ4_K_M6.8 GB84details
RTX 507012 GBFitsQ6_K8.6 GB93details
RTX 508016 GBFitsQ8_010.5 GB103details
2x RTX 30902×24 GBFitsF1618.2 GB53details
2x RTX 40902×24 GBFitsF1618.2 GB57details
2x RTX 50902×32 GBFitsF1618.2 GB102details
RTX A600048 GBFitsF1618.2 GB44details
RTX 6000 Ada48 GBFitsF1618.2 GB55details
RTX PRO 6000 Blackwell96 GBFitsF1618.2 GB102details
L40S48 GBFitsF1618.2 GB49details
A100 40GB40 GBFitsF1618.2 GB88details
A100 80GB80 GBFitsF1618.2 GB116details
H100 PCIe80 GBFitsF1618.2 GB114details
H100 SXM80 GBFitsF1618.2 GB190details
H200141 GBFitsF1618.2 GB273details
RX 76008 GBFitsQ4_K_M6.8 GB54details
RX 7700 XT12 GBFitsQ6_K8.6 GB60details
RX 7800 XT16 GBFitsQ8_010.5 GB67details
RX 7900 GRE16 GBFitsQ8_010.5 GB62details
RX 7900 XT20 GBFitsQ8_010.5 GB86details
RX 907016 GBFitsQ8_010.5 GB68details
RX 9070 XT16 GBFitsQ8_010.5 GB68details
Ryzen AI Max+ 395 128GB128 GBFitsF1618.2 GB15details
Arc A770 16GB16 GBFitsQ8_010.5 GB60details
Mac M1 16GB16 GBFitsQ6_K8.6 GB9.4details
Mac M2 24GB24 GBFitsQ8_010.5 GB11details
Mac M3 24GB24 GBFitsQ8_010.5 GB11details
Mac M4 16GB16 GBFitsQ6_K8.6 GB17details
Mac M4 32GB32 GBFitsF1618.2 GB6.8details
Mac M1 Pro 32GB32 GBFitsF1618.2 GB11details
Mac M3 Pro 36GB36 GBFitsF1618.2 GB8.5details
Mac M4 Pro 24GB24 GBFitsQ8_010.5 GB29details
Mac M4 Pro 64GB64 GBFitsF1618.2 GB16details
Mac M1 Max 64GB64 GBFitsF1618.2 GB23details
Mac M2 Max 96GB96 GBFitsF1618.2 GB23details
Mac M3 Max 128GB128 GBFitsF1618.2 GB23details
Mac M4 Max 64GB64 GBFitsF1618.2 GB31details
Mac M2 Ultra 192GB192 GBFitsF1618.2 GB45details
Mac M3 Ultra 256GB256 GBFitsF1618.2 GB47details

Prefer an API?

IBM: Granite 4.2 8B is on OpenRouter at $0.06 per 1M input tokens and $0.15 per 1M output tokens. Price details and history →

Questions

How much VRAM does granite 4.2 8b need?

About 6.8 GB at 4-bit (Q4_K_M) with an 8K context: 5.0 GB of weights (real GGUF file size), 1.3 GB of KV cache and 0.6 GB of runtime overhead. At 8-bit it needs about 10.5 GB.

What is the smallest GPU that runs granite 4.2 8b?

At Q4_K_M and 8K context, the smallest single consumer GPU in our table that fits it is the RTX 4060 (8 GB).

Can I run granite 4.2 8b on a Mac?

Yes. At Q4_K_M with 8K context it fits on a Mac M1 16GB or larger, using the default macOS GPU memory limit.

Similar models

Same family or a similar parameter count, with the VRAM each needs at Q4_K_M and 8K context.

granite 4.1 8b

8.8B · 6.8 GB at 4-bit

granite 3.3 8b instruct

8.2B · 6.3 GB at 4-bit

granite 3.1 8b instruct

8.2B · 6.3 GB at 4-bit

internlm3 8b instruct

8.8B · 6.0 GB at 4-bit

Qwen3 VL 8B Instruct

8.8B · 6.5 GB at 4-bit

Molmo2 8B

8.7B · 6.5 GB at 4-bit

Model facts

Hugging Face repoibm-granite/granite-4.2-8b
Parameters8,791,592,960 (8.8B)
Layers / hidden size40 / 4096
AttentionGQA; 32 heads, 8 KV heads, head dim 128
KV cache per token (FP16)160.0 KiB
Max context131,072 tokens
Licenseapache-2.0
Released on HFAug 7, 2026
Downloads (30d) / likes59,718 / 127

Sources: Hugging Face model API · config.json (ibm-granite/granite-4.2-8b) · GGUF files · GPU specs from vendor pages (list). Data fetched Nov 17, 2026. Spot an error? Report it.