selfhost.directory

The index of self-hostable projects — ranked, categorized and served with live updates.
Own your data and stop giving it to big tech.

tencent logo

Can the Arc A310 run WeMM-Video-2B?

WeMM-Video-2B·3.3B·Arc A310·4 GB VRAM

Saved in this browser, used on every model page.
Yes

Yes. The Arc A310 runs WeMM-Video-2B fully in VRAM at Q6_K (about 3.5 GB of 4.0 GB), at roughly 25 tokens/s.

best quant
Q6_K
needs
3.5 GB
usable
4.0 GB
est. speed
~25 tok/s

Estimates for an 8K context; offload assumes 32 GB of system RAM.Try other contexts on the WeMM-Video-2B page →

Every quantisation of WeMM-Video-2B on the Arc A310

QuantQualityMemoryFitSpeed
Q2_K Heavy loss1.9 GBfits~63 tok/s
Q3_K_M Noticeable loss2.4 GBfits~42 tok/s
Q4_K_M Small loss2.8 GBfits~34 tok/s
Q5_K_M Near-lossless3.2 GBfits~29 tok/s
Q6_K pickNear-lossless3.5 GBfits~25 tok/s
Q8_0 Lossless4.3 GBoffload~18 tok/s
FP16 / BF16 Lossless7.3 GBoffload~7 tok/s

Speed is a bandwidth-based estimate for generating text (124 GB/s); prompt processing and long contexts are slower.

Measured speed on the Arc A310

No community reports for this pair yet.

Report your own speed →

Vision models that fit the Arc A310 comfortably

All that fit →
ModelParamsBest quantMemorySpeedLicence
3.8BQ6_K≈3.3 GBFits in memory~22 tok/sNon-commercial
LiquidAI logod1-3BLiquidAI
3.1BQ8_0≈3.8 GBFits in memory~21 tok/sConditions
LiquidAI logoLFM2.5-VL-3BLiquidAI
3.1BQ8_0≈3.8 GBFits in memory~21 tok/sConditions
4.3BQ4_K_M≈3.6 GBFits in memory~26 tok/sConditions
3.8BQ5_K_M≈3.5 GBFits in memory~25 tok/sCheck licence
3BQ8_0≈3.8 GBFits in memory~22 tok/sCommercial OK

Other GPUs that run WeMM-Video-2B

Frequently asked questions

Can the Arc A310 run WeMM-Video-2B?
Yes. The Arc A310 runs WeMM-Video-2B fully in VRAM at Q6_K (about 3.5 GB of 4.0 GB), at roughly 25 tokens/s.
How fast does WeMM-Video-2B run on the Arc A310?
An estimate from memory bandwidth gives about 25 tokens/s at Q6_K.
Which quantisation of WeMM-Video-2B should I use on the Arc A310?
Q6_K: the largest one that fits. Expected quality: near-lossless compared with the full-precision weights.

Memory and speed are estimates from the model's size and the card's memory bandwidth; real numbers depend on the runtime, drivers and context length. Community figures are medians of user reports.