BestLLMfor EN Your hardware. Your LLM. Your call.
APIOpen data Find my LLM
Model fiche

Salamandra 7B Instruct

By BSC · Spain

Updated 2026-07-13

chat general multilingual fr
Parameters
7.7B
License
Apache 2.0
Context
8k
VRAM (Q4)
5 GB
Released
February 2025

Overview

Barcelona Supercomputing Center's 7.8B trained on 7.8T tokens covering 35 European languages and 92 programming languages — built for EU sovereignty under Apache 2.0.

When to pick this model

  • EU-sovereign chat deployments
  • Multilingual workloads spanning all 35 EU languages
  • Code assistance across an unusually broad language set
  • Public-sector projects requiring open European provenance
  • Catalan, Occitan, and other low-resource Romance language use cases

VRAM requirements by quantization

VRAM REQUIRED (GB)812Q4_K_M5 GBQ5_K_M6 GBQ8_09 GBFP1616 GB
QuantizationVRAM required
Q4_K_M (recommended)5 GB
Q5_K_M6 GB
Q8_09 GB
FP16 (no quantization)16 GB

VRAM figures include model weights plus a typical 8k KV cache and ~600 MB runtime overhead (Ollama / llama.cpp baseline). Add headroom for higher context lengths.

In practice, Salamandra 7B Instruct fits an 8 GB consumer card at Q4_K_M (5 GB). Stepping up to Q8_0 nearly doubles the footprint to 9 GB, and unquantized FP16 weights take 16 GB — plan your GPU around the Q4 or Q5 figure unless you specifically need the higher fidelity.

Without a GPU, Salamandra 7B Instruct needs roughly 10 GB of system RAM to run on CPU via llama.cpp or Ollama — workable for background jobs, but far slower than GPU inference. Throughput estimates from our compatibility engine: around 10 tokens/sec on entry-level GPUs, on the order of 30 tokens/sec on a mid-range card, and up to 80 tokens/sec on high-end hardware — assuming the chosen quantization fully fits in VRAM.

What hardware do you need

The table below matches Salamandra 7B Instruct to common GPU memory tiers, using the highest-fidelity quantization that fully fits each card class. Spilling layers to system RAM works but costs most of the speed, so size your card to the quantization you actually want to run.

GPU memoryExample cardsBest fit for Salamandra 7B Instruct
8 GBRTX 5070 Laptop, RTX 5060, RTX 5060 Ti 8GBQ5_K_M (6 GB used)
12 GBRTX 5070, RTX 5070 Ti Laptop, RTX 4080 LaptopQ8_0 (9 GB used)
16 GBRTX 5080, RTX 4080 Super, Radeon RX 9070 XTFP16 (16 GB used)
24 GBRTX 4090, Radeon RX 7900 XTX, RTX 5090 LaptopFP16 (16 GB used)
32 GBRTX 5090FP16 (16 GB used)

Which GPU should you buy to run Salamandra 7B Instruct?

To run Salamandra 7B Instruct locally at Q4, you need ~5 GB of VRAM. The best value for this is a RTX 5060 (8 GB VRAM).

Check RTX 5060 price on Amazon →

As an Amazon Associate, BestLLMfor earns from qualifying purchases, at no extra cost to you. It does not influence our independent rankings.

Strengths

  • Backed by EU sovereignty and BSC infrastructure
  • 35 European languages natively supported
  • Apache 2.0 license
  • Coverage of 92 programming languages
  • 7.8T tokens of training data

Limitations

  • 8k context falls short for long-document use
  • No official Ollama distribution
  • Quality trails frontier 7B models on English benchmarks

Typical workloads

In our catalog grid, Salamandra 7B Instruct is filed under European Multilingual, EU Sovereignty — the use cases where its size/quality trade-off makes the most sense. Its tags translate to concrete workloads: multilingual workloads; French-language output where quality matters.

Note the 8k-token context window — fine for short interactions, limiting for long documents or big retrieval contexts. The Apache 2.0 license is permissive, so shipping it inside a commercial product raises no special legal questions.

Architecture & training

Architecture: Dense 7.7B · RoPE · SwiGLU · GQA · 256k vocab

Training: 7.8T tokens, 35 EU languages + 92 programming languages.

Verdict

The reference EU-sovereign 7B — choose it when European language breadth and provenance matter more than top-tier English benchmarks.

Quick start

# HuggingFace : BSC-LT/salamandra-7b-instruct

Or use the open-source MCP server to query this model from Claude Desktop, Cursor, or any MCP-compatible client.

Similar models worth comparing

Frequently asked questions

How much VRAM does Salamandra 7B Instruct need?

At the recommended Q4_K_M quantization, Salamandra 7B Instruct needs about 5 GB of VRAM. Q8_0 takes 9 GB, and unquantized FP16 weights take 16 GB.

Can Salamandra 7B Instruct run without a GPU?

Yes — with roughly 10 GB of system RAM it runs CPU-only through llama.cpp or Ollama. Expect a fraction of GPU speed, which is fine for background or batch jobs but slow for interactive chat.

What context window does Salamandra 7B Instruct support?

Salamandra 7B Instruct supports a 8k-token context window (8,192 tokens).

Can I use Salamandra 7B Instruct commercially?

Yes. Salamandra 7B Instruct is released under Apache 2.0, a permissive open-source license that allows commercial use, modification and redistribution.

How fast is Salamandra 7B Instruct on consumer hardware?

Our compatibility engine estimates on the order of 30 tokens/sec on a mid-range GPU and up to 80 tokens/sec on high-end cards, assuming the quantization fully fits in VRAM.

Which quantization of Salamandra 7B Instruct should I download first?

Start with Q4_K_M (5 GB) — the standard size/quality sweet spot. Step up to Q5_K_M or Q8_0 only if you have VRAM headroom. It already fits an 8 GB card at Q5_K_M.

Tools

Is Salamandra 7B Instruct the right pick for you?

Compute self-hosted ROI → Back to catalog