RunsOnMy ModelsGPUs & MacsAPI pricesVRAM calculatorNew

Can I run internlm2 chat 7b locally?

internlm2 chat 7b by internlm has 7.7B parameters. At 4-bit with an 8K context it needs about 6.0 GB of GPU memory. It fits on 65 of 65 GPUs and Macs we track.

6.0 GBQ4_K_M · 8K context
RTX 4060smallest consumer GPU that fits
M1 16GBsmallest Mac that fits
32Kmax 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 QuantFactory/internlm2-chat-7b-GGUF where available (marked •), otherwise parameter count × bits per weight. How we calculate.

QuantQuality4K ctx8K ctx32K ctx
2-bit (Q2_K) •rough3.9 GB4.4 GB7.4 GB
3-bit (Q3_K_M) •usable4.6 GB5.1 GB8.1 GB
4-bit (Q4_K_M) •good, most popular5.5 GB6.0 GB9.0 GB
5-bit (Q5_K_M) •very good6.2 GB6.7 GB9.7 GB
6-bit (Q6_K) •near lossless7.0 GB7.5 GB10.5 GB
8-bit (Q8_0) •effectively lossless8.7 GB9.2 GB12.2 GB
16-bit (F16/BF16)original weights15.5 GB16.0 GB19.0 GB

Which GPUs and Macs run internlm2 chat 7b?

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 GBFitsQ8_09.2 GB44details
RTX 4060 Ti 16GB16 GBFitsQ8_09.2 GB35details
RTX 309024 GBFitsF1616.0 GB60details
RTX 409024 GBFitsF1616.0 GB65details
RTX 509032 GBFitsF1616.0 GB116details
RTX 5070 Ti16 GBFitsQ8_09.2 GB109details
RX 7900 XTX24 GBFitsF1616.0 GB62details
Mac M4 Pro 48GB48 GBFitsF1616.0 GB18details
Mac M4 Max 128GB128 GBFitsF1616.0 GB35details
Mac M3 Ultra 512GB512 GBFitsF1616.0 GB53details
DGX Spark128 GBFitsF1616.0 GB18details
RTX 407012 GBFitsQ8_09.2 GB61details
RTX 5060 Ti 16GB16 GBFitsQ8_09.2 GB54details
RTX 40608 GBFitsQ5_K_M6.7 GB49details
Arc B58012 GBFitsQ8_09.2 GB55details
RTX 3060 Ti8 GBFitsQ5_K_M6.7 GB81details
RTX 30708 GBFitsQ5_K_M6.7 GB81details
RTX 3080 10GB10 GBFitsQ6_K7.5 GB120details
RTX 3080 12GB12 GBFitsQ8_09.2 GB111details
RTX 3090 Ti24 GBFitsF1616.0 GB65details
RTX 4060 Ti 8GB8 GBFitsQ5_K_M6.7 GB52details
RTX 4070 Super12 GBFitsQ8_09.2 GB61details
RTX 4070 Ti12 GBFitsQ8_09.2 GB61details
RTX 4070 Ti Super16 GBFitsQ8_09.2 GB82details
RTX 408016 GBFitsQ8_09.2 GB87details
RTX 4080 Super16 GBFitsQ8_09.2 GB89details
RTX 50608 GBFitsQ5_K_M6.7 GB81details
RTX 507012 GBFitsQ8_09.2 GB82details
RTX 508016 GBFitsQ8_09.2 GB117details
2x RTX 30902×24 GBFitsF1616.0 GB60details
2x RTX 40902×24 GBFitsF1616.0 GB65details
2x RTX 50902×32 GBFitsF1616.0 GB116details
RTX A600048 GBFitsF1616.0 GB50details
RTX 6000 Ada48 GBFitsF1616.0 GB62details
RTX PRO 6000 Blackwell96 GBFitsF1616.0 GB116details
L40S48 GBFitsF1616.0 GB56details
A100 40GB40 GBFitsF1616.0 GB100details
A100 80GB80 GBFitsF1616.0 GB132details
H100 PCIe80 GBFitsF1616.0 GB129details
H100 SXM80 GBFitsF1616.0 GB216details
H200141 GBFitsF1616.0 GB310details
RX 76008 GBFitsQ5_K_M6.7 GB52details
RX 7700 XT12 GBFitsQ8_09.2 GB53details
RX 7800 XT16 GBFitsQ8_09.2 GB76details
RX 7900 GRE16 GBFitsQ8_09.2 GB70details
RX 7900 XT20 GBFitsF1616.0 GB52details
RX 907016 GBFitsQ8_09.2 GB78details
RX 9070 XT16 GBFitsQ8_09.2 GB78details
Ryzen AI Max+ 395 128GB128 GBFitsF1616.0 GB17details
Arc A770 16GB16 GBFitsQ8_09.2 GB68details
Mac M1 16GB16 GBFitsQ8_09.2 GB8.3details
Mac M2 24GB24 GBFitsQ8_09.2 GB12details
Mac M3 24GB24 GBFitsQ8_09.2 GB12details
Mac M4 16GB16 GBFitsQ8_09.2 GB15details
Mac M4 32GB32 GBFitsF1616.0 GB7.8details
Mac M1 Pro 32GB32 GBFitsF1616.0 GB13details
Mac M3 Pro 36GB36 GBFitsF1616.0 GB9.7details
Mac M4 Pro 24GB24 GBFitsQ8_09.2 GB33details
Mac M4 Pro 64GB64 GBFitsF1616.0 GB18details
Mac M1 Max 64GB64 GBFitsF1616.0 GB26details
Mac M2 Max 96GB96 GBFitsF1616.0 GB26details
Mac M3 Max 128GB128 GBFitsF1616.0 GB26details
Mac M4 Max 64GB64 GBFitsF1616.0 GB35details
Mac M2 Ultra 192GB192 GBFitsF1616.0 GB52details
Mac M3 Ultra 256GB256 GBFitsF1616.0 GB53details

Questions

How much VRAM does internlm2 chat 7b need?

About 6.0 GB at 4-bit (Q4_K_M) with an 8K context: 4.4 GB of weights (real GGUF file size), 1.0 GB of KV cache and 0.6 GB of runtime overhead. At 8-bit it needs about 9.2 GB.

What is the smallest GPU that runs internlm2 chat 7b?

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 internlm2 chat 7b 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.

Model facts

Hugging Face repointernlm/internlm2-chat-7b
Parameters7,737,708,544 (7.7B)
Layers / hidden size32 / 4096
AttentionGQA; 32 heads, 8 KV heads, head dim 128
KV cache per token (FP16)128.0 KiB
Max context32,768 tokens
Licenseother
Released on HFJan 10, 2024
Downloads (30d) / likes49,311 / 83

Sources: Hugging Face model API · config.json (internlm/internlm2-chat-7b) · GGUF files · GPU specs from vendor pages (list). Data fetched Nov 17, 2026. Spot an error? Report it.