Leaderboard / Mistral AI
Devstral Small 2
Mistral AI · Devstral · released 2025-12-09 · data checked 2026-11-01
Devstral Small 2 is a 24B-parameter dense model from Mistral AI (Europe), released 2025-12-09 under the Apache-2.0 license. It doesn't have enough reported benchmarks for an index score yet. It needs about 16.9 GB of memory at 4-bit with an 8K-token context.
Agentic coding model.
Benchmarks
Each bar shows this model's score. The tick marks the best medium model on that benchmark. Sources: 1 reported by Mistral AI, 1 measured by another lab. How we label sources.
Specifications
- Lab
- Mistral AI (FR · Europe)
- Architecture
- Dense transformer
- Total parameters
- 24B
- Active parameters
- All (dense)
- Size class
- Medium
- Context window
- 256K tokens
- Input modalities
- text, image
- Reasoning
- Instruct (no thinking)
- License
- Apache-2.0
- Openness
- Open weights
- Memory to run
- ~16.9 GB at 4-bit, 8K context (more sizes)
- Hugging Face
- 131.3K downloads (30 days) · 713 likes
Compare with
- Devstral Small 2 vs LFM2.5 24B-A2B (24B, index 61.4)
- Devstral Small 2 vs Magistral Small 1.2 (24B, index 55.9)
- Devstral Small 2 vs Mistral Small 3.2 (24B, index 38.6)
Sources
Something wrong here? Report it.
Run it locally
Uses unsloth/Devstral-Small-2-24B-Instruct-2512-GGUF (Q4_K_M). Swap the tag for another size: BF16, IQ4_NL, IQ4_XS, Q2_K, Q2_K_L, Q3_K_M, Q3_K_S, Q4_0….
ollama run hf.co/unsloth/Devstral-Small-2-24B-Instruct-2512-GGUF:Q4_K_Mllama-server -hf unsloth/Devstral-Small-2-24B-Instruct-2512-GGUF:Q4_K_Mpip install -U mlx-lm
mlx_lm.chat --model mlx-community/Devstral-Small-2-24B-Instruct-2512-4bitvllm serve mistralai/Devstral-Small-2-24B-Instruct-2512Memory needed: weights, context cache and runtime buffers.
| Quantization | 4K context | 32K context | 256K context |
|---|---|---|---|
| 16-bit F16 | 50 GB | 54 GB | 91 GB |
| 8-bit Q8_0 | 27 GB | 32 GB | 68 GB |
| 4-bit Q4_K_M | 16 GB | 21 GB | 57 GB |