command-a-vision-07-2025
Coheredocs.cohere.com/docs/models
Command A Vision is our first model capable of processing images, excelling in enterprise use cases such as analyzing charts, graphs, and diagrams, table understanding, OCR, document Q\&A, and object detection. It officially supports English, Portuguese, Italian, French, German,
Updated 29 min ago · first seen 11 Sept 2026
model_01M294H95KTBNNVCCG8P0CE7Z0
- Parameters
- 111.9B
- T2 · 2 h ago
- Context
- 128K tokens
- T1 · 2 h ago
- Released
- 28 Jul 2025
- T2 · 2 h ago
- License
- cc-by-nc-4.0
- T2 · 2 h ago
Specification
- Family
- Command
Source:Cohere — docs & blogT1observed 2 h agohigh
- Release date
- 28 Jul 2025
Source:Hugging Face Hub (public pages, model cards, papers)T2observed 2 h agomedium
- Status
- active
Source:Cohere — docs & blogT1observed 2 h agohigh
- Openness
- restricted
Source:Hugging Face Hub (public pages, model cards, papers)T2observed 2 h agomedium
- License
- cc-by-nc-4.0
Source:Hugging Face Hub (public pages, model cards, papers)T2observed 2 h agomedium
- Architecture
- Cohere2VisionForConditionalGeneration
Source:Hugging Face Hub (public pages, model cards, papers)T2observed 2 h agomedium
- Parameters
- 111.9B
Source:Hugging Face Hub (public pages, model cards, papers)T2observed 2 h agomedium
- Context window
- 128K tokens
Source:Cohere — docs & blogT1observed 2 h agohigh
- Max output
- 8K tokens
Source:Cohere — docs & blogT1observed 2 h agohigh
- Modalities
- text, image
Source:Cohere — docs & blogT1observed 2 h agohigh
- Input modalities
- text, image
Source:Cohere — docs & blogT1observed 2 h agohigh
- Languages
- en, fr, de, es, it, pt
Source:Hugging Face Hub (public pages, model cards, papers)T2observed 2 h agomedium
- API model id
- command-a-vision-07-2025
Source:Cohere — docs & blogT1observed 2 h agohigh
- Base model
- CohereLabs/c4ai-command-a-03-2025, google/siglip2-so400m-patch16-512
Source:Hugging Face Hub (public pages, model cards, papers)T2observed 2 h agomedium
- File size
- 223.7 GB
Source:Hugging Face Hub (public pages, model cards, papers)T2observed 2 h agomedium
- Hugging Face repo
- CohereLabs/command-a-vision-07-2025
Source:Hugging Face Hub (public pages, model cards, papers)T2observed 2 h agomedium
- Pipeline tag
- image-text-to-text
Source:Hugging Face Hub (public pages, model cards, papers)T2observed 2 h agomedium
- Official page
Source:Cohere — docs & blogT1observed 2 h agohigh
- Model card
Source:Hugging Face Hub (public pages, model cards, papers)T2observed 2 h agomedium
- Downloads
- 17,970
Source:Hugging Face Hub (public pages, model cards, papers)T2observed 29 min agomedium
- Likes
- 88
Source:Hugging Face Hub (public pages, model cards, papers)T2observed 29 min agomedium
- Endpoints
- Chat
Source:Cohere — docs & blogT1observed 2 h agohigh
- Gated
- auto
Source:Hugging Face Hub (public pages, model cards, papers)T2observed 2 h agomedium
- Hf inference providers
- cohere
Source:Hugging Face Hub (public pages, model cards, papers)T2observed 2 h agomedium
- Last modified
- 2025-10-30T06:38:41+00:00
Source:Hugging Face Hub (public pages, model cards, papers)T2observed 2 h agomedium
- Library name
- transformers
Source:Hugging Face Hub (public pages, model cards, papers)T2observed 2 h agomedium
- Downloads all time
- 678,849
Source:Hugging Face Hub (public pages, model cards, papers)T2observed 29 min agomedium
- Model type
- cohere2_vision
Source:Hugging Face Hub (public pages, model cards, papers)T2observed 2 h agomedium
- Tags
- transformers, safetensors, cohere2_vision, image-text-to-text, en, fr, de, es, it, pt
Source:Hugging Face Hub (public pages, model cards, papers)T2observed 2 h agomedium
- Weights dtype
- F16
Source:Hugging Face Hub (public pages, model cards, papers)T2observed 2 h agomedium
Each value shows its source, tier and observation time. Conflicting claims are kept side by side and flagged — never averaged. How AI Atlas records facts →
Provenance
Attributed facts
32
Source tiers
T1T211 / 21
Freshest observation
29 min ago
Conflicts
None
Modalities
- Modalities
- textimage
- Input
- textimage
Capabilities
Tool calling
Unavailable
Structured output
Unavailable
Reasoning
Unavailable
Vision
Yes
Cohere — docs & blog · T1
Audio
Unavailable
Fine-tuning available
Unavailable
- Context window
- 128K tokens
Source:Cohere — docs & blogT1observed 2 h agohigh
- Max output
- 8K tokens
Source:Cohere — docs & blogT1observed 2 h agohigh
- Languages
- en, fr, de, es, it, pt
Source:Hugging Face Hub (public pages, model cards, papers)T2observed 2 h agomedium
No benchmark results recorded
Current prices
No current prices recorded
Available through 1
Price history
Memory need = bytes per parameter (4-bit ≈ 0.5 × 1.15 overhead, 8-bit 1.0, fp16 2.0) + a KV-cache allowance. Not a measurement.
| Hardware | Quantization | Memory | Est. need | Fits |
|---|---|---|---|---|
| NVIDIA DGX B200 | 4bit | 1,440 GB | 64.8 GB | Yes |
| Apple M3 Ultra | 4bit | — | 64.8 GB | Yes |
| Mac Studio (Apple M5 Ultra) | 4bit | — | 64.8 GB | Yes |
| AMD Instinct MI325X | 4bit | 256 GB | 64.8 GB | Yes |
| AMD Instinct MI300X | 4bit | 192 GB | 64.8 GB | Yes |
| Apple M2 Ultra | 4bit | — | 64.8 GB | Yes |
| NVIDIA B200 | 4bit | 180 GB | 64.8 GB | Yes |
| NVIDIA H200 | 4bit | 141 GB | 64.8 GB | Yes |
| NVIDIA H200 NVL | 4bit | 141 GB | 64.8 GB | Yes |
| Apple M1 Ultra | 4bit | — | 64.8 GB | Yes |
| Apple M3 Max | 4bit | — | 64.8 GB | Yes |
| Apple M4 Max | 4bit | — | 64.8 GB | Yes |
| Mac Studio (Apple M5 Max) | 4bit | — | 64.8 GB | Yes |
| MacBook Pro (Apple M5 Max) | 4bit | — | 64.8 GB | Yes |
| NVIDIA DGX Spark | 4bit | 128 GB | 64.8 GB | Yes |
| Apple M2 Max | 4bit | — | 64.8 GB | Yes |
| NVIDIA H100 NVL | 4bit | 94 GB | 64.8 GB | Yes |
| NVIDIA A100 80GB | 4bit | 80 GB | 64.8 GB | Yes |
| NVIDIA H100 SXM | 4bit | 80 GB | 64.8 GB | Yes |
| Apple M1 Max | 4bit | — | 64.8 GB | No |
| Apple M4 Pro | 4bit | — | 64.8 GB | No |
| Mac mini (Apple M5 Pro) | 4bit | — | 64.8 GB | No |
| MacBook Pro (Apple M5 Pro) | 4bit | — | 64.8 GB | No |
| Apple M3 Pro | 4bit | — | 64.8 GB | No |
| Apple M1 Pro | 4bit | — | 64.8 GB | No |
| Apple M2 Pro | 4bit | — | 64.8 GB | No |
| Apple M4 | 4bit | — | 64.8 GB | No |
| iMac (Apple M4) | 4bit | — | 64.8 GB | No |
| Mac mini (Apple M6) | 4bit | — | 64.8 GB | No |
| MacBook Air (Apple M5) | 4bit | — | 64.8 GB | No |
| MacBook Pro (Apple M5) | 4bit | — | 64.8 GB | No |
| NVIDIA GeForce RTX 5090 | 4bit | 32 GB | 64.8 GB | No |
| Apple M2 | 4bit | — | 64.8 GB | No |
| Apple M3 | 4bit | — | 64.8 GB | No |
| NVIDIA GeForce RTX 3090 | 4bit | 24 GB | 64.8 GB | No |
| NVIDIA GeForce RTX 4090 | 4bit | 24 GB | 64.8 GB | No |
| Apple M1 | 4bit | — | 64.8 GB | No |
Ancestors 2
- c4ai-command-a-03-2025111.1B
- siglip2-so400m-patch16-512Google
This model
command-a-vision-07-2025
111.9B params
Quantizations 0
None recorded.
Descendants 0
None recorded.
Papers 0
No papers linked yet.
Repositories 0
No repositories linked yet.
As of
Rewind the record: see this entity's attributes exactly as AI Atlas knew them on a given day.
Claim history · Architecture
Architecturearchitecture1
| Value | Valid from → to | Status | Source | Confidence | Extractor |
|---|---|---|---|---|---|
| Cohere2VisionForConditionalGeneration | → current | current | Hugging Face Hub (public pages, model cards, papers)T2 | medium | deterministic |
Claims are temporal and append-only: a new observation closes the previous claim (valid_to) instead of overwriting it. Conflicting claims from different sources are kept side by side and flagged — never averaged. Methodology →
| Source | Document | Type | Tier | Last observed | Snapshots |
|---|---|---|---|---|---|
| Cohere — docs & blog | docs.cohere.com/docs/models.md | model_docs | T1· Official | 51 min ago | 1 |
| Hugging Face Hub (public pages, model cards, papers) | huggingface.co/CohereLabs/command-a-vision-07-2025 | model_page | T2· Quality secondary | 29 min ago | 2 |
| Hugging Face Hub (public pages, model cards, papers) | huggingface.co/models?author=CohereLabs&p=0&sort=downloads | listing | T2· Quality secondary | 42 min ago | 3 |
Tier 1 = official/primary, 2 = quality secondary, 3 = community, 4 = unverified. Every snapshot is archived; see all sources and the methodology.