aya-expanse-8b
CohereLabs CC BY-NC 4.0 out 2 years ago
8.03B parameters, text in, text out.
Verified 2026-10-25 Rank #601 in LLMs · score 74.1 method v1 Sources 3 Downloads 30d 17.4k · #467 of 619 tracked
Where it's used
Hugging Face does not say who downloads a model or with what. These are the proxies: API providers listing it, quantized builds, fine-tunes/adapters/merges, and Spaces that load it.
No API provider serves it, so its downloads are people running it themselves: 53 quantized builds and 86 Spaces.
- As a service0no API provider listed
- On own hardware53quantized builds
- Built on11384 fine-tunes · 29 adapters · 0 merges
- In apps86Hugging Face Spaces
Top Spaces running it: pliny-the-prompter/obliteratus ♥ 621 · vilarin/aya-expanse-8b ♥ 67 · feel-fl/open-human-feedback-chat ♥ 14 · build-small-hackathon/multi-agent-lab ♥ 6
Will it fit?
Weights only, an estimate: leave room for the KV cache. Method
Fits a 16 GB card at INT4
About 4.8 GB at INT4, 9.6 GB at FP8, 19 GB at BF16: weights only, so leave room for the KV cache.
Estimate params × bytes per param × 1.2 · how we estimate
What people built on it
166 derivatives on the Hub. Which build to run, by runtime: the most downloaded in each class.
166 derivatives on the Hub, most of them fine-tunes
From the Hugging Face model tree; quantized builds in teal, fine-tunes, merges and adapters in greys.
Reference
Everything else, folded: the full tables, every source.
Use-case evidence
3 use cases
What it is good at 3 use cases
Labels are evidence, not verdicts: measured a third-party leaderboard row for this model; author-reported a number the model author published in the card or eval files; structural a deterministic signal from the name, config or chat template, or a decision-model estimate. Each chip links to its source. All use cases.
Facts
16 fields
Listed in LLMs
Facts
- aya-expanse-8b has 8.03B parameters. (safetensors metadata, observed 2026-10-24)
- License: CC BY-NC 4.0 (friction: non-commercial). Non-commercial only; commercial hosting not permitted without a separate license. (license file, observed 2026-10-24)
- Created 2024-10-23, last modified 2026-01-09. (HF API, observed 2026-10-24)
- Access is gated on Hugging Face (auto). (HF API, observed 2026-10-24)
| Pipeline tag | text-generation |
|---|---|
| Library | transformers |
| Architectures | CohereForCausalLM |
| Model type | cohere |
| Params by dtype | F16: 8,028,033,024 |
| Size band | M |
| MoE | no |
| Gated | auto |
| Papers | arXiv:2408.14960, arXiv:2407.02552, arXiv:2406.18682, arXiv:2410.10801, arXiv:2412.04261 |
| License | CC BY-NC 4.0 (cc-by-nc-4.0) |
| Hub snapshot | 2026-10-25 |
Servability
vLLM · SGLang
Servability
vLLM: supported · CohereForCausalLM · 7050ef881ff0835f6890a47db1859a24c65481d6SGLang: supported · CohereForCausalLM · 773cb938fe9484a8f7e7e46fb1b1fe96412c59ed
vLLM registry (7050ef881ff0835f6890a47db1859a24c65481d6) · SGLang registry (773cb938fe9484a8f7e7e46fb1b1fe96412c59ed) · Matched by architecture class name against each project's registry on main; 'supported' means the architecture is registered, not that this checkpoint was tested.
Hardware fit
3 estimates
Hardware fit ESTIMATE
Weights-only memory for aya-expanse-8b: params × bytes/param × 1.2 overhead. KV cache and activations are extra. How this is computed.
| BF16 (2 bytes/param) | 19 GB · Single GPU (<24 GB) |
|---|---|
| FP8 (1 byte/param) | 9.6 GB · Single GPU (<24 GB) |
| INT4 (0.5 bytes/param) | 4.8 GB · Single GPU (<24 GB) |
Weights only (params × bytes/param × 1.2 overhead); KV cache and activations extra.
All variants
40 variants
Lineage
quantized: 53finetune: 84adapter: 29
Top variants (40 shown, ranked by downloads)
| Variant | Relation | Quant | Params | Band | License | Author | Likes | Downloads |
|---|---|---|---|---|---|---|---|---|
| Orion-zhen/aya-expanse-8b-AWQ | quantized | AWQ | 9.08B | L | gpl-3.0 | Orion-zhen | 1 | 204.8k |
| bartowski/aya-expanse-8b-GGUF | quantized | GGUF | — | — | cc-by-nc-4.0 | bartowski | 47 | 136k |
| lmstudio-community/aya-expanse-8b-GGUF | quantized | GGUF | — | — | cc-by-nc-4.0 | lmstudio-community | 6 | 13.1k |
| lenML/aya-expanse-8b-abliterated | finetune | — | 8.03B | M | cc-by-nc-4.0 | lenML | 10 | 8.1k |
| matrixportalx/aya-expanse-8b-GGUF | quantized | GGUF | — | — | cc-by-nc-4.0 | matrixportalx | 1 | 7.6k |
| DevQuasar-3/CohereForAI.aya-expanse-8b-GGUF | quantized | GGUF | — | — | — | DevQuasar-3 | 3 | 5.5k |
| prithivMLmods/Aya-Expanse-8B-GGUF | quantized | GGUF | — | — | apache-2.0 | prithivMLmods | 5 | 5.1k |
| tensorblock/aya-expanse-8b-GGUF | quantized | GGUF | — | — | cc-by-nc-4.0 | tensorblock | 3 | 4.8k |
| matrixportalx/Aya-Medikal-GGUF | quantized | GGUF | — | — | cc-by-nc-4.0 | matrixportalx | 0 | 4.9k |
| Yousefbahr/Turjman-Cold-Start | finetune | — | 8.03B | M | — | Yousefbahr | 2 | 3.2k |
| kevinbazira/aya-expanse-8b-awq-4bit | quantized | awq | 8.03B | M | cc-by-nc-4.0 | kevinbazira | 0 | 2.8k |
| mlx-community/aya-expanse-8b-4bit | quantized | -4bit | 8.03B | M | cc-by-nc-4.0 | mlx-community | 0 | 2.3k |
| cstr/aya-expanse-8b-Q4_K_M-GGUF | quantized | GGUF | — | — | cc-by-nc-4.0 | cstr | 0 | 1.6k |
| Aya168/aya-expanse-8b-tofu-13lang-v3 | finetune | — | 8.03B | M | cc-by-nc-4.0 | Aya168 | 1 | 1.1k |
| lang-uk/OmniGEC-Minimal-8B | finetune | — | 8.02B | M | mit | lang-uk | 3 | 1k |
| WueNLP/centurio_aya | finetune | — | 8.48B | M | cc-by-nc-4.0 | WueNLP | 4 | 739 |
| huggingkot/aya-expanse-8b-bnb-4bit | finetune | -bnb | 8.03B | M | — | huggingkot | 0 | 830 |
| itlwas/aya-expanse-8b-Q4_K_M-GGUF | quantized | GGUF | — | — | cc-by-nc-4.0 | itlwas | 0 | 749 |
| NikolayKozloff/aya-expanse-8b-Q8_0-GGUF | quantized | Q8_0 | — | — | cc-by-nc-4.0 | NikolayKozloff | 1 | 608 |
| mlx-community/aya-expanse-8b-4bit-mlx | quantized | -4bit | 8.03B | M | cc-by-nc-4.0 | mlx-community | 0 | 633 |
| mlx-community/aya-expanse-8b-8bit | quantized | -8bit | 8.03B | M | — | mlx-community | 0 | 548 |
| matrixportalx/lenML-aya-expanse-8b-abliterated-GGUF | quantized | GGUF | — | — | cc-by-nc-4.0 | matrixportalx | 2 | 424 |
| YorkieOH10/aya-expanse-8b-Q8_0-GGUF | quantized | Q8_0 | — | — | cc-by-nc-4.0 | YorkieOH10 | 0 | 509 |
| doruoros/Aya-Hermes-13B | finetune | — | 11.3B | L | — | doruoros | 0 | 490 |
| Svngoku/French-Aya-Expanse-8B | finetune | — | 8.03B | M | — | Svngoku | 1 | 370 |
| matrixportalx/Aya-Empati-v3 | finetune | — | 8.03B | M | cc-by-nc-4.0 | matrixportalx | 4 | 191 |
| Iatalking/aya-expanse-8b-Q4_K_M-GGUF | quantized | GGUF | — | — | cc-by-nc-4.0 | Iatalking | 0 | 357 |
| matrixportalx/aya-expanse-8b-Q3_K_S-GGUF | quantized | GGUF | — | — | cc-by-nc-4.0 | matrixportalx | 0 | 333 |
| alabenayed/TounsiLM-8b | adapter | — | — | — | cc-by-nc-4.0 | alabenayed | 0 | 305 |
| Cossale/aya-expanse-8b-formal | finetune | — | 8.03B | M | — | Cossale | 0 | 294 |
| Lucy-in-the-Sky/aya-expanse-8b-Q8_0-GGUF | quantized | Q8_0 | — | — | cc-by-nc-4.0 | Lucy-in-the-Sky | 0 | 285 |
| zeuvue/aya-expanse-8b-Q5_K_M-GGUF | quantized | GGUF | — | — | cc-by-nc-4.0 | zeuvue | 0 | 280 |
| thanhh12/aya-expanse-8b-Q4_0-GGUF | quantized | Q4_0 | — | — | cc-by-nc-4.0 | thanhh12 | 0 | 273 |
| jeanflop/aya-expanse-8b-Q4_K_M-GGUF | quantized | GGUF | — | — | cc-by-nc-4.0 | jeanflop | 1 | 222 |
| matrixportalx/Aya-Empati | finetune | — | 8.03B | M | cc-by-nc-4.0 | matrixportalx | 1 | 218 |
| matrixportalx/aya-expanse-8b-Q2_K-GGUF | quantized | Q2_K | — | — | cc-by-nc-4.0 | matrixportalx | 0 | 258 |
| zeuvue/aya-expanse-8b-IQ4_XS-GGUF | quantized | GGUF | — | — | cc-by-nc-4.0 | zeuvue | 0 | 226 |
| lang-uk/OmniGEC-Fluency-8B | finetune | — | 8.02B | M | mit | lang-uk | 2 | 124 |
| matrixportalx/Aya-Empati-v2 | finetune | — | 8.03B | M | cc-by-nc-4.0 | matrixportalx | 1 | 171 |
| LingoIITGN/COMI-LINGUA-LID | adapter | — | — | — | apache-2.0 | LingoIITGN | 0 | 214 |
Papers
5 papers
Papers 5
arXiv links found in the model card and its tags. "Peer-reviewed" is shown only with evidence: a published venue in OpenAlex (measured) or the authors' own journal reference on arXiv (reported).
- Aakanksha, Arash Ahmadian, Seraphina Goldfarb-Tarrant +3 · 2024-10-14 · PDF · HF paper page (4 upvotes)
Similar models
3 models
Similar models
- Llama-3.1-8B-Instruct · 8.03B same size band (M), same task (text-generation), more providers
- Qwen3-8B · 8.19B same size band (M), same task (text-generation), more providers, permissive license
- Meta-Llama-3-8B-Instruct · 8.03B same size band (M), same task (text-generation), more providers
Sources
3 sources
Sources
Hub snapshot: 2026-10-25 · HF API: 2026-10-24 21:40 UTC · Providers: 2026-10-24 21:44 UTC · Articles: 2026-10-24 21:46 UTC · Summary: 2026-10-07 15:11 UTC · Last verified 2026-10-25
Follow aya-expanse-8b
Watch this model
- /feeds/models/CohereLabs--aya-expanse-8b.xml(RSS 2.0) · every change to aya-expanse-8b
- /feeds/families/CohereLabs--aya-expanse.xml(RSS 2.0) · every change across the aya-expanse family
Email alerts are not open yet; the feeds above are the same data. See feeds for the item schema.