NVIDIA / discrete

NVIDIA B300

270 GB decides what fits. 7700 GB/s decides how fast it runs once it does.

Quick answer

The NVIDIA B300 fits 285 of 320 sized models entirely in its 270 GB; the largest widely used one is MiniMax-M3 (264.0 GB at Q4_K_M). Memory decides what fits; its 7700 GB/s of bandwidth decides how fast it answers.

Largest popular model that fitsMiniMax-M3427B params · Q4_K_M · needs 264.0 GB~324 tok/s Faster than you readOpen in the calculator →
Best fast pickGigaChat3.5-432B-A28B-Reasoning438B params · Q4_K_M · needs 270.0 GB~429 tok/s Faster than you readOpen in the calculator →
Most downloaded that fitsGLM-5.3-Flash321B params · Q4_K_M · needs 198.3 GB~554 tok/s Faster than you readOpen in the calculator →

At 8,192 tokens of context with the whole model in device memory. Speeds are estimates from memory bandwidth, not benchmarks run on this card.

NVIDIA B300 · 270 GB · 7700 GB/s · Q4_K_M where published · 8,192 tokensSpecificationpublished dataDecode speedestimate

285 of 320 fit entirely

285 of the 320 models the engine can size fit entirely in device memory at Q4_K_M where it is published, otherwise the nearest published format, and 8,192 tokens. 2 run with some layers on system memory (32 GB assumed), and 33 do not run at all.

89%of the models the engine can size fit entirely

270GB of device memory

7700GB/s memory bandwidth

2run with system memory

33do not run at 8,192 tokens

Featured models · Q4_K_M at 8,192 tokens, including offload

Model	Needs	Verdict	Decode	Calculator
Qwen3.5-2B2.3B parameters	2.32 GB	fits	~4245 tok/s	Open →
NVIDIA-Nemotron-3-Nano-4B-BF164.0B parameters	3.46 GB	fits	~1967 tok/s	Open →
Qwen3-VL-8B-Instruct8.8B parameters	7.04 GB	fits	~890 tok/s	Open →
Qwen3.5-9B9.7B parameters	7.05 GB	fits	~1077 tok/s	Open →
gemma-4-12B-it12B parameters	9.54 GB	fits	~641 tok/s	Open →
gpt-oss-20b21B parameters	13.8 GB	fits	~2093 tok/s	Open →
Qwen3.8-27B28B parameters	18.6 GB	fits	~354 tok/s	Open →
Qwen3.6-27B28B parameters	18.6 GB	fits	~354 tok/s	Open →
Kimi-Linear-48B-A3B-Instruct49B parameters	30.4 GB	fits	~2801 tok/s	Open →
Qwen3.8-Flash-Next180B parameters	111.6 GB	fits	~1455 tok/s	Open →
DeepSeek-V4-Flash-Vision-Exp305B parameters	188.1 GB	fits	~465 tok/s	Open →
GLM-5.3-Flash321B parameters	198.3 GB	fits	~554 tok/s	Open →

These examples are selected from prominent labs using the catalogue’s latest Hugging Face 30-day downloads and repository-creation freshness signal, with newer releases guaranteed a place. Offloaded rows assume 32 GB of system RAM, and a speed is only shown for a row that runs. Fitting in memory is not the same as loading: whether the runtime and version you have supports each architecture and format on this machine has not been tested here. Each “Open” link carries the same model, format, context and RAM into the calculator.

Fits entirely in NVIDIA B300 memory — 285 of 320 sized models

Every model that fitswhole model resident · offload off

The most demanding model that fits is GigaChat3.5-432B-A28B-Reasoning at Q4_K_M: 270.0 GB of the 270 GB, leaving 0.04 GB spare.

Fully resident at 8,192 tokens (or the model’s own maximum, where that is shorter), offload off, at 7700 GB/s. Each row is a run of the engine for this configuration; the rows start with current, prominent releases and “fits” is memory, not a tested runtime. Older or less prominent models remain available through this search and “Show all”.

Find a model in this list

Showing 40 of 285 models that fit.

Model	Format	Needs	Spare	Decode	Downloads / 30d	Calculator
GLM-5.3-FlashNEWzai-org · 321B params	Q4_K_M	198.3 GB	71.7 GB	~554 tok/s	5.4M	Open →
Qwen3.8-27BQwen · 28B params	Q4_K_M	18.6 GB	251.4 GB	~354 tok/s	6.9M	Open →
DeepSeek-V4-Flash-0731DeepSeek · 304B params	Q4_K_M	187.5 GB	82.5 GB	~465 tok/s	4.5M	Open →
Qwen3.8-Flash-NextNEWQwen · 180B params	Q4_K_M	111.6 GB	158.4 GB	~1455 tok/s	1.4M	Open →
gemma-4-26B-A4B-itGoogle · 26B params	Q4_K_M	16.8 GB	253.2 GB	~1967 tok/s	13M	Open →
DeepSeek-V4-Flash-Vision-ExpNEWDeepSeek · 305B params	Q4_K_M	188.1 GB	81.9 GB	~465 tok/s	915.3K	Open →
gemma-4-31B-itGoogle · 31B params	Q4_K_M	22.2 GB	247.8 GB	~254 tok/s	9.9M	Open →
Qwen3.5-9BQwen · 9.7B params	Q4_K_M	7.05 GB	263.0 GB	~1077 tok/s	9M	Open →
Qwen3.5-4BQwen · 4.7B params	Q4_K_M	3.98 GB	266.0 GB	~1939 tok/s	7.8M	Open →
Qwen3.6-35B-A3BQwen · 36B params	Q4_K_M	23.1 GB	246.9 GB	~2824 tok/s	3.3M	Open →
gemma-4-12B-itGoogle · 12B params	Q4_K_M	9.54 GB	260.5 GB	~641 tok/s	1.9M	Open →
Inkling-Smallthinkingmachines · 266B params	Q4_K_M	165.5 GB	104.5 GB	~735 tok/s	657.4K	Open →
Qwen3.6-27BQwen · 28B params	Q4_K_M	18.6 GB	251.4 GB	~354 tok/s	2.5M	Open →
Qwen3.5-2BQwen · 2.3B params	Q4_K_M	2.32 GB	267.7 GB	~4245 tok/s	4.9M	Open →
gemma-4-E4B-itGoogle · 8.0B params	Q4_K_M	5.86 GB	264.1 GB	~1731 tok/s	4.4M	Open →
NVIDIA-Nemotron-3.5-Lightning-30B-A3B-BF16NVIDIA · 32B params	Q4_K_M	20.3 GB	249.7 GB	~268 tok/s	530K	Open →
NVIDIA-Nemotron-3-Nano-4B-BF16NVIDIA · 4.0B params	Q4_K_M	3.46 GB	266.5 GB	~1967 tok/s	3.4M	Open →
gemma-4-E2B-itGoogle · 5.1B params	Q4_K_M	4.02 GB	266.0 GB	~4572 tok/s	3M	Open →
Qwen3.5-0.8BQwen · 873M params	Q4_K_M	1.46 GB	268.5 GB	~9299 tok/s	2.6M	Open →
DeepSeek-V4-FlashDeepSeek · 284B params	Q4_K_M	175.9 GB	94.1 GB	~465 tok/s	1.1M	Open →
Qwen3-VL-8B-InstructQwen · 8.8B params	Q4_K_M	7.04 GB	263.0 GB	~890 tok/s	14.6M	Open →
MiniMax-M2.7MiniMaxAI · 229B params	Q4_K_M	141.2 GB	128.8 GB	~609 tok/s	1.1M	Open →
Qwen3.5-27BQwen · 28B params	Q4_K_M	18.6 GB	251.4 GB	~354 tok/s	1.9M	Open →
North-Micro-Vision-InstructCohereLabs · 2.5B params	Q4_K_M	2.99 GB	267.0 GB	~2791 tok/s	180.7K	Open →
Hy3Tencent · 299B params	Q4_K_M	181.2 GB	88.8 GB	~371 tok/s	304.9K	Open →
Qwen3.5-35B-A3BQwen · 36B params	Q4_K_M	23.1 GB	246.9 GB	~2824 tok/s	1.6M	Open →
NVIDIA-Nemotron-3-Super-120B-A12B-BF16NVIDIA · 124B params	Q4_K_M	76.9 GB	193.1 GB	~69 tok/s	1.2M	Open →
granite-4.2-8bIBM · 8.8B params	Q4_K_M	7.49 GB	262.5 GB	~810 tok/s	128.4K	Open →
GLM-4.7-Flashzai-org · 31B params	Q4_K_M	20.4 GB	249.6 GB	~267 tok/s	1.8M	Open →
granite-4.1-3bIBM · 3.4B params	Q4_K_M	3.72 GB	266.3 GB	~1890 tok/s	521.8K	Open →
LFM2.5-2.6BLiquidAI · 2.7B params	Q4_K_M	2.61 GB	267.4 GB	~3032 tok/s	108.1K	Open →
MiMo-V2.6-Flash-RLNEWXiaomiMiMo · 309B params	Q4_K_M	192.2 GB	77.8 GB	~540 tok/s	48.6K	Open →
MiniMax-M3MiniMaxAI · 427B params	Q4_K_M	264.0 GB	6.04 GB	~324 tok/s	177.2K	Open →
MiMo-V2.5XiaomiMiMo · 311B params	Q4_K_M	192.2 GB	77.8 GB	~540 tok/s	245.2K	Open →
Qwen3-VL-4B-InstructQwen · 4.4B params	Q4_K_M	4.72 GB	265.3 GB	~1414 tok/s	3.5M	Open →
granite-4.1-30bIBM · 29B params	Q4_K_M	20.4 GB	249.6 GB	~266 tok/s	301.5K	Open →
Qwen3.5-122B-A10BQwen · 125B params	Q4_K_M	77.9 GB	192.1 GB	~1034 tok/s	512.6K	Open →
Qwen3-VL-2B-InstructQwen · 2.1B params	Q4_K_M	3.05 GB	267.0 GB	~2563 tok/s	2.8M	Open →
granite-4.2-3bIBM · 3.7B params	Q4_K_M	3.72 GB	266.3 GB	~1890 tok/s	50.6K	Open →
LFM2.5-230MLiquidAI · 230M params	Q4_K_M	1.05 GB	268.9 GB	~21132 tok/s	88.6K	Open →

Sized at the format most people actually download, not at FP16. The catalogue ordered by downloads →

Rent it instead

Rent an B300 by the hour

What the NVIDIA B300 costs on Vast.ai and RunPod right now, per machine, from one card to eight. Every price is read from the provider's own API.

Prices · 18:30 UTC, 22 Sept

1× B300 · 270 GB

$7.89per hour

RunPod · Secure CloudRent on RunPod

2× B300 · 540 GB

$15.78per hour

RunPod · Secure CloudRent on RunPod

Buy one or rent one?

Your price, your hours, your electricity. Everything else is arithmetic.

Price you would pay ($)

Resale after 3 years ($)

Hours of use per day4 h

$ per kWh

Watts

Rent $/h

Watts start at the NVIDIA B300’s published board power (1100 W), an upper bound: decoding rarely holds a card at its limit. Rent starts at the cheapest live price (RunPod). The electricity price is a placeholder — put yours in.

At 4 h a day, renting costs $11,519.40 a year.

That is $959.95 a month, and nothing when the machine is stopped.
Owning would cost $401.50 a year in electricity, on top of the price.
Every $1,000 of purchase price takes 0.1 years of this use to earn back.

Enter the price you would pay to see the exact break-even.

On-demand prices for the whole machine. Vast hosts below 98% measured reliability are left out; RunPod’s Community Cloud is vetted third-party hosts and its Secure Cloud is data-centre capacity. Every rentable card compared.

Referral links Vast.ai, RunPod and Novita pay us a share of what you spend if you sign up through these buttons. It costs you nothing, and it never decides an order or a recommendation: both are computed from the live price and the speed, and options that pay us nothing are listed and recommended on the same terms. How we rank

Under the hood

The specification behind every figure

What the manufacturer publishes for this device, and the pages it was read from.

Manufacturer specification

Memory scope	dedicated
Capacity	270 GB
Published options	270 GB
Bandwidth	7700 GB/s
Memory type	HBM3e
Bus width	Not published
FP32 peak	Not published
Dense matrix peak	Not published without sparsity
TDP	1100 W

Source ledger

NVIDIA exemplar-performance README (reference architectures), commit 82e952b173b2 ↗

Retrieved 2026-09-14

Caveats

This is the B300 GPU as configured in NVIDIA's 8-GPU B300 reference architecture. It is not the GB300 Grace-Blackwell rack platform, whose CPU and GPU memory tiers the engine does not model.
The source is a repository file that changes over time; it is cited at a pinned commit.
No dense matrix throughput is published for this device in a form this catalogue accepts, so no compute roof is priced and time to first token is withheld.

Published capacity is a hardware ceiling, not guaranteed free runtime memory. The calculator shows the runtime reserve separately rather than folding it into a single number.

Run the diagnostic on the NVIDIA B300 →