llmrun
CohereCommand

C4ai Command A 03 2025 — Hardware Requirements & GPU Compatibility

ChatFunctions

Command A (command-a-03-2025) is Cohere and Cohere Labs' 111.1-billion-parameter dense (non-mixture-of-experts) language model tuned for enterprise agentic work: tool use, retrieval-augmented generation, and multilingual tasks across 23 languages. Its architecture repeats a pattern of three sliding-window attention layers (4,096-token window) followed by one global-attention layer without positional embeddings, a design later reused and scaled up in the larger Command A+. At release, Cohere reported roughly 150% higher throughput than its predecessor Command R+ 08-2024 and said the model could run on as few as two GPUs. At 111 billion parameters, it needs a multi-GPU server to run. Context length is 256,000 tokens. It is released under the CC BY-NC 4.0 license, restricting use to non-commercial purposes only. It was published in March 2025.

111.1B params · denseCC BY-NC 4.01.7K downloads397 likes3.2K quant downloads

C4ai Command A 03 2025 runs on a 96 GB+ Mac or a workstation GPU

Best local pick → Mac Studio (M3 Ultra, 96GB) · Q4_K_M · 78.7 GB · ~8 tok/s

Macs from48 GB IQ2_S · ~11 tok/s

Score 79.8#83 of 100 open#44 of 60 that fit ≤96 GBCC BY-NC 4.0

Not set up for it?

Reference GPU · RTX 4090 24 GB

Estimates, ±20% · how we estimate

llmrun Composite Score →

79.8

Benchmarks · tier chart ↓
Among 100 open modelsHow the Composite Score is built

Run C4ai Command A 03 2025

One recommended way. Every other method, OS and quant is in the step-by-step guide.

Other ways

each opens step-by-step

Run in cloud

Fits on RTX PRO 6000 (96 GB) (17 GB headroom) · Q4_K_M

Generation speed
~14 tok/s
generation speed
Cost per 1M output tokens
$30.05
per 1M output tokens
Deployon Vast.aiBest optionfrom $1.54/hr (opens in new tab)
Compare GPUs →
or

How Much VRAM Does C4ai Command A 03 2025 Need?

Select a quantization level; the listed hardware adheres to it.

QuantizationVRAM+32KOn RTX 4090 (reference)
51.9 GB63.4 GB2-bit · K-quant — 2-bit quantization with K-quant improvementsDoes not fit: needs 57.3 GB
53.5 GB64.9 GB3-bit small — 3-bit small quantizationDoes not fit: needs 58.8 GB
59.5 GB71 GB3-bit medium — 3-bit medium quantizationDoes not fit: needs 64.9 GB
61.1 GB72.6 GB4-bit · legacy — 4-bit legacy quantizationDoes not fit: needs 66.4 GB
73.3 GB84.8 GB4-bit medium — 4-bit medium quantization — most popular sweet spotDoes not fit: needs 78.7 GB
87.0 GB98.5 GB5-bit medium — 5-bit medium quantization — good quality/size tradeoffDoes not fit: needs 92.4 GB
100.8 GB112.3 GB6-bit, very good — 6-bit quantization, very good qualityDoes not fit: needs 106.1 GB
122.2 GB133.7 GB8-bit, near-lossless — 8-bit quantization, near-losslessDoes not fit: needs 127.5 GB

Selected Q4_K_M

4-bit medium quantization

78.7GBat 16K context

Weights
72.2 GB
KV cache
6.1 GB
Overhead
0.3 GB

Doesn't fit RTX 4090 (reference) · needs 54.7 GB more

Or rent RTX PRO 6000 from $1.54/hr.

Which Hardware Can Run C4ai Command A 03 2025?

Q4_K_M · 73.3 GB

C4ai Command A 03 2025 (Q4_K_M) requires 73.3 GB of VRAM to load the model weights. For comfortable inference with headroom for KV cache and system overhead, 96+ GB is recommended. No consumer GPU has enough memory. 30 devices with unified memory can run C4ai Command A 03 2025, including Mac Studio (M3 Ultra, 256GB), Mac Studio (M3 Ultra, 512GB), Mac Studio (M3 Ultra, 96GB).

By VRAM tier · typical GPUs and Macs at Q4_K_M
8–16 GB · RTX 4060, RTX 3060 12GB, RTX 5060 Ti 16GBDoesn't fit — smallest quant needs 39 GB8 GB tier →12 GB tier →16 GB tier →
24 GB · RTX 4090referenceDoesn't fit — smallest quant needs 39 GB24 GB tier →
Mac 32 GB · Mac Mini M4Doesn't fit — smallest quant needs 39 GBMacs →
Mac 64 GB · Mac Studio M4 MaxIQ3_M · 60.3 GB · ~7 tok/s · 16KMacs →

How good is C4ai Command A 03 2025?

The llmrun Score merges 3 public benchmark boards into a single figure, where 100 represents the average for open models.

Score vs VRAM in the ≤96 GB tier

  • C4ai Command A 03 2025
  • Other models
  • Best Score for its VRAM

#44 of 60 that fit ≤96 GB

Each dot is a model at the best quant that fits ≤96 GB. The line joins the models with the best Score for their VRAM; a dot below it is beaten by a model that needs no more memory.

Full ≤96 GB ranking →

Benchmarks

Here are all 4 boards that have results for this model, sorted by their weight in the llmrun Score; badges show who ran the test.

  • LMArena Text

    LMArena · Oct 2026Arena

    1353.7

    #73/187 open

    Arena
  • Aider Polyglot

    Epoch AI · Mar 2025Independent

    12.0%

    #16/18 open

    Independent
  • DTBench

    Epoch AI · Mar 2025Lab-reported

    61.3%

    #43/68 open

    Lab-reported
  • LMCA

    Epoch AI · Mar 2025Lab-reported

    10.3%

    #47/54 open

    Lab-reported
All 4 C4ai Command A 03 2025 benchmarks

In the ≤96 GB class

60 models fit ≤96 GB

Same Score, only models that fit ≤96 GB at their best quant: the ranking a ≤96 GB owner actually chooses from.

  1. #1

    MiMo V2.6 Flash RLXiaomi

    131.9

    ~153 tok/s · 93.8 GB

  2. #2

    DeepSeek V4 Flash 0731DeepSeek

    127.6

    ~103 tok/s · 95.5 GB

  3. #3

    Qwen3.8 27BAlibaba

    122.1

    ~67 tok/s · 17.9 GB

  4. #4

    Inkling SmallThinking Machines

    119.6

    ~356 tok/s · 92.9 GB

  5. #5

    MiniMax M2.5MiniMax

    111.8

    ~138 tok/s · 93.9 GB

  6. #6

    Qwen3.6 27BAlibaba

    111.2

    ~67 tok/s · 17.9 GB

Full ≤96 GB ranking →RTX PRO 6000 Blackwell Workstation Edition page →

Alternatives to C4ai Command A 03 2025

Official models with an llmrun Score, each answering one question; numbers are at the best quant that fits a 24 GB GPU.

Similar models

10
ModelScoreVRAMtok/sWhy it is similar
MiniMax M2.5MiniMax111.8139.6 GB—needs > 24 GBsimilar score
Qwen3.5 122B A10BAlibaba109.475.5 GB—needs > 24 GBsimilar score
MiniMax M2.7MiniMax107.4139.6 GB—needs > 24 GBsimilar score
MiniMax M2.1MiniMax106.2139.6 GB—needs > 24 GBsimilar score
MiniMax M2MiniMax104.0139.6 GB—needs > 24 GBsimilar score
Qwen3 235B A22BAlibaba99.6142.9 GB—needs > 24 GBsimilar score
Phi 4Microsoft83.912.5 GB~69similar score
Qwen2.5 72B InstructAlibaba81.749.3 GB—needs > 24 GBsimilar score
Llama 3.3 70B InstructMeta79.751.3 GB—needs > 24 GBsimilar score
Gemma 3 27B ITGoogle79.421.7 GB~36similar score

Specifications

Publisher
Cohere
Family
Command
Parameters
111.1B
Release Date
2025-03-11
License
CC BY-NC 4.0

Resources

Frequently Asked Questions

How much VRAM does C4ai Command A 03 2025 need?

C4ai Command A 03 2025 requires 73.3 GB of VRAM at Q4_K_M, or 244.3 GB at BF16.

VRAM = Weights + KV Cache + Overhead

Weights = 111.1B × 4.8 bits ÷ 8 = 66.6 GB

KV Cache + Overhead ≈ 6.7 GB (at 2K context + ~0.3 GB framework)

Fit ratings and hardware model lists check this model with room for a 16K-token context, which needs a little more memory.

VRAM usage by quantization

73.3 GB

Learn more about VRAM estimation →

Can NVIDIA GeForce RTX 5090 run C4ai Command A 03 2025?

No — C4ai Command A 03 2025 requires at least 33.6 GB at IQ2_XXS, which exceeds the NVIDIA GeForce RTX 5090's 32 GB of VRAM.

What's the best quantization for C4ai Command A 03 2025?

For C4ai Command A 03 2025, Q4_K_M (73.3 GB) offers the best balance of quality and VRAM usage. Q4_K_L (74.8 GB) provides better quality if you have the VRAM. The smallest option is IQ2_XXS at 33.6 GB.

VRAM requirement by quantization

IQ2_XXS
33.6 GB
IQ3_XS
50.4 GB
Q4_0
61.1 GB
Q4_K_M ★
73.3 GB
Q4_K_L
74.8 GB
BF16
244.3 GB

★ Recommended — best balance of quality and VRAM usage.

Learn more about quantization →

Can I run C4ai Command A 03 2025 on a Mac?

Yes — Mac Studio M5 Max (48 GB) and 55 other Macs can run C4ai Command A 03 2025, the smallest at IQ2_S. Apple Silicon uses unified memory, so the model shares RAM with the system. At Q4_K_M you need at least 73.3 GB of usable unified memory (RAM minus macOS overhead), from Mac Studio (M3 Ultra, 96GB).

Can I run C4ai Command A 03 2025 locally?

Yes — C4ai Command A 03 2025 can run locally on consumer hardware. Popular tools include Ollama, LM Studio, and llama.cpp.

How fast is C4ai Command A 03 2025?

At Q4_K_M, C4ai Command A 03 2025 can reach ~66 tok/s on AMD Instinct MI350X. Speed depends mainly on GPU memory bandwidth. Real-world results typically within ±20%.

tok/s = (bandwidth GB/s ÷ model GB) × efficiency

Example: NVIDIA B200 → 8000 ÷ 73.3 × 0.65 = ~71 tok/s

Estimated speed at Q4_K_M (73.3 GB)

~71 tok/s
~71 tok/s
~66 tok/s

Real-world results typically within ±20%. Speed depends on batch size, quantization kernel, and software stack.

Learn more about tok/s estimation →

What's the download size of C4ai Command A 03 2025?

At Q4_K_M, the download is about 66.63 GB. The full-precision BF16 version is 222.12 GB. The smallest option (IQ2_XXS) is 30.54 GB.

Which GPUs can run C4ai Command A 03 2025?

No single consumer GPU has enough VRAM to run C4ai Command A 03 2025 at Q4_K_M (73.3 GB). Multi-GPU or professional hardware is required.

Which devices can run C4ai Command A 03 2025?

30 devices with unified memory can run C4ai Command A 03 2025 at Q4_K_M (73.3 GB), including ASUS Ascent GX10, Beelink GTR9 Pro (Ryzen AI Max+ 395, 128 GB), Framework Desktop (Ryzen AI Max+ 395, 128 GB), GMKtec EVO-X2 (Ryzen AI Max+ 395, 128 GB). Apple Silicon Macs use unified memory shared between CPU and GPU, making them well-suited for local LLM inference.