MiniCPM5-1B

OpenBMB · openbmb/minicpm5-1b

ProviderOpenBMB
Familyminicpm
Typellm-chat
Statusactive
Released2026-05-21
Updated2026-08-17
Parameters1.1B
Open weightsyes

What it fits on

Ordered by predicted speed. Bandwidth sets decode rate; memory decides whether it runs at all.

Every figure here is computed, not measured. Fit is weights at each quantisation against device memory, with a 25% allowance for the KV cache, activations and the OS. Decode rate is the memory-bandwidth roofline at 70% efficiency. Nobody has run this model on these devices.
DeviceMemoryBandwidthBest qualityWeightstok/sFastest
Cerebras WSE-344.0 GB21000000.0 GB/sbf162.16 GB~6801570.1~27206280.6 q4
NVIDIA GB200 Grace Blackwell Superchip372.0 GB16000.0 GB/sbf162.16 GB~5182.1~20728.6 q4
AMD Instinct MI355X288.0 GB8000.0 GB/sbf162.16 GB~2591.1~10364.3 q4
NVIDIA B200180.0 GB7700.0 GB/sbf162.16 GB~2493.9~9975.6 q4
Google TPU7x (Ironwood)192.0 GB7380.0 GB/sbf162.16 GB~2390.3~9561.1 q4
AMD Instinct MI325X256.0 GB6000.0 GB/sbf162.16 GB~1943.3~7773.2 q4
AMD Instinct MI300X192.0 GB5300.0 GB/sbf162.16 GB~1716.6~6866.3 q4
NVIDIA H200 SXM141.0 GB4800.0 GB/sbf162.16 GB~1554.6~6218.6 q4
NVIDIA H100 NVL94.0 GB3900.0 GB/sbf162.16 GB~1263.1~5052.6 q4
NVIDIA H100 SXM80.0 GB3350.0 GB/sbf162.16 GB~1085.0~4340.0 q4
AMD Instinct MI250X128.0 GB3200.0 GB/sbf162.16 GB~1036.4~4145.7 q4
Google TPU v5p95.0 GB2765.0 GB/sbf162.16 GB~895.5~3582.2 q4
NVIDIA A100 80GB SXM80.0 GB2039.0 GB/sbf162.16 GB~660.4~2641.6 q4
NVIDIA H100 PCIe80.0 GB2000.0 GB/sbf162.16 GB~647.8~2591.1 q4
NVIDIA GeForce RTX 509032.0 GB1792.0 GB/sbf162.16 GB~580.4~2321.6 q4
Google TPU v6e (Trillium)32.0 GB1638.0 GB/sbf162.16 GB~530.5~2122.1 q4
AMD Instinct MI21064.0 GB1600.0 GB/sbf162.16 GB~518.2~2072.9 q4
NVIDIA A100 40GB SXM40.0 GB1555.0 GB/sbf162.16 GB~503.6~2014.6 q4
Google TPU v432.0 GB1200.0 GB/sbf162.16 GB~388.7~1554.6 q4
NVIDIA GeForce RTX 409024.0 GB1008.0 GB/sbf162.16 GB~326.5~1305.9 q4
AMD Radeon RX 7900 XTX24.0 GB960.0 GB/sbf162.16 GB~310.9~1243.7 q4
NVIDIA GeForce RTX 508016.0 GB960.0 GB/sbf162.16 GB~310.9~1243.7 q4
NVIDIA GeForce RTX 309024.0 GB936.0 GB/sbf162.16 GB~303.2~1212.6 q4
NVIDIA L40S48.0 GB864.0 GB/sbf162.16 GB~279.8~1119.3 q4
Apple M3 Ultra512.0 GB819.0 GB/sbf162.16 GB~265.3~1061.0 q4
AMD Radeon RX 7900 XT20.0 GB800.0 GB/sbf162.16 GB~259.1~1036.4 q4
Apple M2 Ultra192.0 GB800.0 GB/sbf162.16 GB~259.1~1036.4 q4
Google TPU v5e16.0 GB800.0 GB/sbf162.16 GB~259.1~1036.4 q4
NVIDIA GeForce RTX 4080 SUPER16.0 GB736.0 GB/sbf162.16 GB~238.4~953.5 q4
NVIDIA GeForce RTX 4070 Ti SUPER16.0 GB672.0 GB/sbf162.16 GB~217.7~870.6 q4
AMD Radeon RX 9070 XT16.0 GB640.0 GB/sbf162.16 GB~207.3~829.1 q4
Apple M4 Max128.0 GB546.0 GB/sbf162.16 GB~176.8~707.4 q4
Apple M1 Max64.0 GB400.0 GB/sbf162.16 GB~129.6~518.2 q4
Apple M2 Max96.0 GB400.0 GB/sbf162.16 GB~129.6~518.2 q4
Apple M3 Max128.0 GB400.0 GB/sbf162.16 GB~129.6~518.2 q4
NVIDIA GeForce RTX 3060 12GB12.0 GB360.0 GB/sbf162.16 GB~116.6~466.4 q4
NVIDIA L424.0 GB300.0 GB/sbf162.16 GB~97.2~388.7 q4
NVIDIA GeForce RTX 4060 Ti 16GB16.0 GB288.0 GB/sbf162.16 GB~93.3~373.1 q4
Apple M4 Pro64.0 GB273.0 GB/sbf162.16 GB~88.4~353.7 q4
Apple M432.0 GB120.0 GB/sbf162.16 GB~38.9~155.5 q4

Where it runs

Taken from the card's availability section.

PlatformId on that platform
Hugging Faceopenbmb/MiniCPM5-1B

Verified benchmark evidence

Each row was checked against its source by a reviewer, and carries the model identifier as actually evaluated — which is not always the same as this card's.
BenchmarkEvaluated asScoreEvidence dateSource kind
aa_lcrMiniCPM5-1B6.0%2026-09-10 evaluatedindependent evaluatorsource
critptMiniCPM5-1B0.0%2026-09-10 evaluatedindependent evaluatorsource
gpqa_diamondMiniCPM5-1B27.78%2026-09-10 evaluatedindependent evaluatorsource

Reported benchmark scores

This card reports no benchmark scores yet.

BenchmarkCatalogue standingScore

Data

This card as JSON · See it in the graph · Edit on GitHub