Leaderboard / Barcelona Supercomputing Center (ALIA)
ALIA 40B Instruct
Barcelona Supercomputing Center (ALIA) · ALIA · released 2026-06-30 · data checked 2026-11-01
ALIA 40B Instruct is a 40.4B-parameter dense model from Barcelona Supercomputing Center (ALIA) (Europe), released 2026-06-30 under the Apache-2.0 license. It doesn't have enough reported benchmarks for an index score yet. It needs about 27.4 GB of memory at 4-bit with an 8K-token context.
Spain's public model, focused on Spanish, Catalan, Basque, Galician and English.
Benchmarks
Each bar shows this model's score. The tick marks the best large model on that benchmark.
No benchmark results for this release yet. Know of one? Send it.
Specifications
- Lab
- Barcelona Supercomputing Center (ALIA) (ES · Europe)
- Architecture
- Dense transformer
- Total parameters
- 40.4B
- Active parameters
- All (dense)
- Size class
- Large
- Context window
- 164K tokens
- Input modalities
- text
- Reasoning
- Instruct (no thinking)
- License
- Apache-2.0
- Openness
- Open weights
- Memory to run
- ~27.4 GB at 4-bit, 8K context (more sizes)
- Hugging Face
- 2.7K downloads (30 days) · 4 likes
Compare with
- ALIA 40B Instruct vs Seed-OSS 36B Instruct (36.2B, index 60.3)
- ALIA 40B Instruct vs Hermes 4.3 36B (36.2B, index 51.4)
- ALIA 40B Instruct vs MiniCPM-V 4.7 35B-A3B (35.2B, index 54.5)
Sources
Something wrong here? Report it.
Run it locally
Uses BSC-LT/ALIA-40b-instruct-2606-GGUF (Q8_0). Swap the tag for another size: Q8_0.
ollama run hf.co/BSC-LT/ALIA-40b-instruct-2606-GGUF:Q8_0llama-server -hf BSC-LT/ALIA-40b-instruct-2606-GGUF:Q8_0vllm serve BSC-LT/ALIA-40b-instruct-2606Memory needed: weights, context cache and runtime buffers.
| Quantization | 4K context | 32K context | 164K context |
|---|---|---|---|
| 16-bit F16 | 83 GB | 89 GB | 114 GB |
| 8-bit Q8_0 | 45 GB | 51 GB | 76 GB |
| 4-bit Q4_K_M | 27 GB | 32 GB | 58 GB |