cohere-transcribe-arabic-07-2026
Coheredocs.cohere.com/docs/models
Finetune optimized for Arabic audio inputs
Updated 23 min ago · first seen 11 Sept 2026
model_01M294H99YT5GCHS2VEGK95FP5
- Parameters
- 2.07B
- T2 · 26 min ago
Specification
- Family
- Audio
Source:Cohere — docs & blogT1observed 33 min agohigh
- Status
- active
Source:Cohere — docs & blogT1observed 33 min agohigh
- Openness
- restricted
Source:Hugging Face Hub (public pages, model cards, papers)T2observed 26 min agomedium
- Parameters
- 2.07B
Source:Hugging Face Hub (public pages, model cards, papers)T2observed 26 min agomedium
- API model id
- cohere-transcribe-arabic-07-2026
Source:Cohere — docs & blogT1observed 33 min agohigh
- Hugging Face repo
- CohereLabs/cohere-transcribe-arabic-07-2026
Source:Hugging Face Hub (public pages, model cards, papers)T2observed 26 min agomedium
- Pipeline tag
- automatic-speech-recognition
Source:Hugging Face Hub (public pages, model cards, papers)T2observed 26 min agomedium
- Official page
Source:Cohere — docs & blogT1observed 33 min agohigh
- Model card
Source:Hugging Face Hub (public pages, model cards, papers)T2observed 26 min agomedium
- Downloads
- 47,352
Source:Hugging Face Hub (public pages, model cards, papers)T2observed 23 min agomedium
- Likes
- 188
Source:Hugging Face Hub (public pages, model cards, papers)T2observed 23 min agomedium
- Endpoints
- Audio Transcriptions
Source:Cohere — docs & blogT1observed 33 min agohigh
- Last modified
- 2026-07-13T16:11:20+00:00
Source:Hugging Face Hub (public pages, model cards, papers)T2observed 26 min 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
14
Source tiers
T1T26 / 8
Freshest observation
23 min ago
Conflicts
None
Modalities
Modalities unavailable.
Capabilities
Tool calling
Unavailable
Structured output
Unavailable
Reasoning
Unavailable
Vision
Unavailable
Audio
Unavailable
Fine-tuning available
Unavailable
No capability flags have been observed from a source yet — we do not infer them.
No structured attributes yet.
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 |
|---|---|---|---|---|
| Apple M3 Ultra | 4bit | — | 1.7 GB | Yes |
| AMD Instinct MI325X | 4bit | 256 GB | 1.7 GB | Yes |
| AMD Instinct MI300X | 4bit | 192 GB | 1.7 GB | Yes |
| Apple M2 Ultra | 4bit | — | 1.7 GB | Yes |
| NVIDIA B200 | 4bit | 180 GB | 1.7 GB | Yes |
| NVIDIA H200 | 4bit | 141 GB | 1.7 GB | Yes |
| Apple M1 Ultra | 4bit | — | 1.7 GB | Yes |
| Apple M3 Max | 4bit | — | 1.7 GB | Yes |
| Apple M4 Max | 4bit | — | 1.7 GB | Yes |
| NVIDIA DGX Spark | 4bit | 128 GB | 1.7 GB | Yes |
| Apple M2 Max | 4bit | — | 1.7 GB | Yes |
| NVIDIA A100 80GB | 4bit | 80 GB | 1.7 GB | Yes |
| NVIDIA H100 SXM | 4bit | 80 GB | 1.7 GB | Yes |
| Apple M1 Max | 4bit | — | 1.7 GB | Yes |
| Apple M4 Pro | 4bit | — | 1.7 GB | Yes |
| Apple M3 Pro | 4bit | — | 1.7 GB | Yes |
| Apple M1 Pro | 4bit | — | 1.7 GB | Yes |
| Apple M2 Pro | 4bit | — | 1.7 GB | Yes |
| Apple M4 | 4bit | — | 1.7 GB | Yes |
| NVIDIA GeForce RTX 5090 | 4bit | 32 GB | 1.7 GB | Yes |
| Apple M2 | 4bit | — | 1.7 GB | Yes |
| Apple M3 | 4bit | — | 1.7 GB | Yes |
| NVIDIA GeForce RTX 3090 | 4bit | 24 GB | 1.7 GB | Yes |
| NVIDIA GeForce RTX 4090 | 4bit | 24 GB | 1.7 GB | Yes |
| Apple M1 | 4bit | — | 1.7 GB | Yes |
No lineage 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 · Parameters
Parametersparameter_count1
| Value | Valid from → to | Status | Source | Confidence | Extractor |
|---|---|---|---|---|---|
| 2.07B | → 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 →
New model: cohere-transcribe-arabic-07-2026 (Cohere)
cohere
| Source | Document | Type | Tier | Last observed | Snapshots |
|---|---|---|---|---|---|
| Cohere — docs & blog | docs.cohere.com/docs/models.md | model_docs | T1· Official | 33 min ago | 1 |
| Hugging Face Hub (public pages, model cards, papers) | huggingface.co/models?author=CohereLabs&p=0&sort=downloads | listing | T2· Quality secondary | 23 min ago | 2 |
Tier 1 = official/primary, 2 = quality secondary, 3 = community, 4 = unverified. Every snapshot is archived; see all sources and the methodology.