tiny-aya-global
Coheredocs.cohere.com/docs/models
Tiny Aya Global is a 3.35B instruction-tuned multilingual model with the best balance across languages and regions. Supports 70 languages.
Updated 5 min ago · first seen 11 Sept 2026
model_01M294H9A6MD8XR7FS8WAHSK2R
- Parameters
- 3.35B
- T1 · 32 min ago
- Context
- 8K tokens
- T1 · 32 min ago
- Released
- 13 Feb 2026
- T2 · 5 min ago
- License
- cc-by-nc-4.0
- T2 · 5 min ago
Specification
- Family
- Aya
Source:Cohere — docs & blogT1observed 32 min agohigh
- Release date
- 13 Feb 2026
Source:Hugging Face Hub (public pages, model cards, papers)T2observed 5 min agomedium
- Status
- active
Source:Cohere — docs & blogT1observed 32 min agohigh
- Openness
- restricted
Source:Hugging Face Hub (public pages, model cards, papers)T2observed 25 min agomedium
- License
- cc-by-nc-4.0
Source:Hugging Face Hub (public pages, model cards, papers)T2observed 5 min agomedium
- Architecture
- Cohere2ForCausalLM
Source:Hugging Face Hub (public pages, model cards, papers)T2observed 5 min agomedium
- Parameters
- 3.35B
Source:Cohere — docs & blogT1observed 32 min agohigh
- Context window
- 8K tokens
Source:Cohere — docs & blogT1observed 32 min agohigh
- Max output
- 8K tokens
Source:Cohere — docs & blogT1observed 32 min agohigh
- Modalities
- text
Source:Cohere — docs & blogT1observed 32 min agohigh
- Input modalities
- text
Source:Cohere — docs & blogT1observed 32 min agohigh
- Languages
- en, nl, fr, it, pt, ro, es, cs, pl, uk, ru, el, de, da, sv, no, ca, gl, cy, ga, eu, hr, lv, lt, sk, sl, et, fi, hu, sr, bg, ar, fa, ur, tr, mt, he, hi, mr, bn, gu, pa, ta, te, ne, tl, ms, id, vi, jv, km, th, lo, zh, my, ja, ko, am, ha, ig, mg, sn, sw, wo, xh, yo, zu
Source:Hugging Face Hub (public pages, model cards, papers)T2observed 5 min agomedium
- API model id
- tiny-aya-global
Source:Cohere — docs & blogT1observed 32 min agohigh
- Base model
- CohereLabs/tiny-aya-base
Source:Hugging Face Hub (public pages, model cards, papers)T2observed 5 min agomedium
- File size
- 6.7 GB
Source:Hugging Face Hub (public pages, model cards, papers)T2observed 5 min agomedium
- Hugging Face repo
- CohereLabs/tiny-aya-global
Source:Hugging Face Hub (public pages, model cards, papers)T2observed 25 min agomedium
- Pipeline tag
- text-generation
Source:Hugging Face Hub (public pages, model cards, papers)T2observed 25 min agomedium
- Official page
Source:Cohere — docs & blogT1observed 32 min agohigh
- Model card
Source:Hugging Face Hub (public pages, model cards, papers)T2observed 25 min agomedium
- Downloads
- 6,966
Source:Hugging Face Hub (public pages, model cards, papers)T2observed 5 min agomedium
- Likes
- 172
Source:Hugging Face Hub (public pages, model cards, papers)T2observed 5 min agomedium
- Endpoints
- Chat
Source:Cohere — docs & blogT1observed 32 min agohigh
- Gated
- auto
Source:Hugging Face Hub (public pages, model cards, papers)T2observed 5 min agomedium
- Hf inference providers
- cohere
Source:Hugging Face Hub (public pages, model cards, papers)T2observed 5 min agomedium
- Last modified
- 2026-07-07T14:58:13+00:00
Source:Hugging Face Hub (public pages, model cards, papers)T2observed 25 min agomedium
- Library name
- transformers
Source:Hugging Face Hub (public pages, model cards, papers)T2observed 5 min agomedium
- Downloads all time
- 116,257
Source:Hugging Face Hub (public pages, model cards, papers)T2observed 5 min agomedium
- Model type
- cohere2
Source:Hugging Face Hub (public pages, model cards, papers)T2observed 5 min agomedium
- Tags
- transformers, safetensors, cohere2, text-generation, en, nl, fr, it, pt, ro, es, cs, pl, uk, ru, el, de, da, sv, no, ca, gl, cy, ga, eu, hr, lv, lt, sk, sl
Source:Hugging Face Hub (public pages, model cards, papers)T2observed 5 min agomedium
- Weights dtype
- BF16
Source:Hugging Face Hub (public pages, model cards, papers)T2observed 5 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
32
Source tiers
T1T212 / 20
Freshest observation
5 min ago
Conflicts
3 flagged
Modalities
- Modalities
- text
- Input
- text
Capabilities
Tool calling
Unavailable
Structured output
Unavailable
Reasoning
Unavailable
Vision
No
Cohere — docs & blog · T1
Audio
Unavailable
Fine-tuning available
Unavailable
- Context window
- 8K tokens
Source:Cohere — docs & blogT1observed 32 min agohigh
- Max output
- 8K tokens
Source:Cohere — docs & blogT1observed 32 min agohigh
- Languages
- en, nl, fr, it, pt, ro, es, cs, pl, uk, ru, el, de, da, sv, no, ca, gl, cy, ga, eu, hr, lv, lt, sk, sl, et, fi, hu, sr, bg, ar, fa, ur, tr, mt, he, hi, mr, bn, gu, pa, ta, te, ne, tl, ms, id, vi, jv, km, th, lo, zh, my, ja, ko, am, ha, ig, mg, sn, sw, wo, xh, yo, zu
Source:Hugging Face Hub (public pages, model cards, papers)T2observed 5 min 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 |
|---|---|---|---|---|
| Apple M3 Ultra | 4bit | — | 2.4 GB | Yes |
| AMD Instinct MI325X | 4bit | 256 GB | 2.4 GB | Yes |
| AMD Instinct MI300X | 4bit | 192 GB | 2.4 GB | Yes |
| Apple M2 Ultra | 4bit | — | 2.4 GB | Yes |
| NVIDIA B200 | 4bit | 180 GB | 2.4 GB | Yes |
| NVIDIA H200 | 4bit | 141 GB | 2.4 GB | Yes |
| Apple M1 Ultra | 4bit | — | 2.4 GB | Yes |
| Apple M3 Max | 4bit | — | 2.4 GB | Yes |
| Apple M4 Max | 4bit | — | 2.4 GB | Yes |
| NVIDIA DGX Spark | 4bit | 128 GB | 2.4 GB | Yes |
| Apple M2 Max | 4bit | — | 2.4 GB | Yes |
| NVIDIA A100 80GB | 4bit | 80 GB | 2.4 GB | Yes |
| NVIDIA H100 SXM | 4bit | 80 GB | 2.4 GB | Yes |
| Apple M1 Max | 4bit | — | 2.4 GB | Yes |
| Apple M4 Pro | 4bit | — | 2.4 GB | Yes |
| Apple M3 Pro | 4bit | — | 2.4 GB | Yes |
| Apple M1 Pro | 4bit | — | 2.4 GB | Yes |
| Apple M2 Pro | 4bit | — | 2.4 GB | Yes |
| Apple M4 | 4bit | — | 2.4 GB | Yes |
| NVIDIA GeForce RTX 5090 | 4bit | 32 GB | 2.4 GB | Yes |
| Apple M2 | 4bit | — | 2.4 GB | Yes |
| Apple M3 | 4bit | — | 2.4 GB | Yes |
| NVIDIA GeForce RTX 3090 | 4bit | 24 GB | 2.4 GB | Yes |
| NVIDIA GeForce RTX 4090 | 4bit | 24 GB | 2.4 GB | Yes |
| Apple M1 | 4bit | — | 2.4 GB | Yes |
Ancestors 1
- tiny-aya-base3.35B
This model
tiny-aya-global
3.35B params
Quantizations 0
None recorded.
Descendants 0
None recorded.
Papers 1
- arXiv:2603.11510Active
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 · Modalities
Modalitiesmodalities1
| Value | Valid from → to | Status | Source | Confidence | Extractor |
|---|---|---|---|---|---|
| text | → current | current | Cohere — docs & blogT1 | high | 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 | 32 min ago | 1 |
| Hugging Face Hub (public pages, model cards, papers) | huggingface.co/CohereLabs/tiny-aya-global | model_page | T2· Quality secondary | 5 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.