Quick start
ollama run llama-proAvailable sizes
| Tag | Size | Quantization | Context | Min RAM |
|---|---|---|---|---|
| llama-pro:8b-instruct-q2_K | 3.5GB | q2_k | 4K context | 4.4 GB |
| llama-pro:8b-text-q2_K | 3.5GB | q2_k | 4K | 4.4 GB |
| llama-pro:8b-instruct-q3_K_S | 3.6GB | q3_k_s | 4K context | 4.5 GB |
| llama-pro:8b-text-q3_K_S | 3.6GB | q3_k_s | 4K | 4.5 GB |
| llama-pro:8b-instruct-q3_K_M | 4.1GB | q3_k_m | 4K context | 5.1 GB |
| llama-pro:8b-text-q3_K_M | 4.1GB | q3_k_m | 4K | 5.1 GB |
| llama-pro:8b-instruct-q3_K_L | 4.5GB | q3_k_l | 4K context | 5.6 GB |
| llama-pro:8b-text-q3_K_L | 4.5GB | q3_k_l | 4K | 5.6 GB |
| llama-pro:latest | 4.7GB | q4_k_m | 4K context | 5.9 GB |
| llama-pro:8b-instruct-q4_K_S | 4.8GB | q4_k_s | 4K context | 6 GB |
| llama-pro:8b-text-q4_K_S | 4.8GB | q4_k_s | 4K | 6 GB |
| llama-pro:8b-instruct-q4_K_M | 5.1GB | q4_k_m | 4K context | 6.4 GB |
| llama-pro:8b-text-q4_K_M | 5.1GB | q4_k_m | 4K | 6.4 GB |
| llama-pro:8b-instruct-q4_1 | 5.3GB | q4_1 | 4K context | 6.6 GB |
| llama-pro:8b-text-q4_1 | 5.3GB | q4_1 | 4K | 6.6 GB |
| llama-pro:8b-instruct-q5_0 | 5.8GB | q5_0 | 4K context | 7.2 GB |
| llama-pro:8b-instruct-q5_K_S | 5.8GB | q5_k_s | 4K | 7.2 GB |
| llama-pro:8b-instruct-q5_K_M | 5.9GB | q5_k_m | 4K context | 7.4 GB |
| llama-pro:8b-text-q5_K_M | 5.9GB | q5_k_m | 4K | 7.4 GB |
| llama-pro:8b-instruct-q5_1 | 6.3GB | q5_1 | 4K context | 7.9 GB |
| llama-pro:8b-text-q5_1 | 6.3GB | q5_1 | 4K | 7.9 GB |
| llama-pro:8b-instruct-q6_K | 6.9GB | q6_k | 4K context | 8.6 GB |
| llama-pro:8b-text-q6_K | 6.9GB | q6_k | 4K | 8.6 GB |
| llama-pro:8b-instruct-q8_0 | 8.9GB | q8_0 | 4K context | 11.1 GB |
| llama-pro:8b-text-q8_0 | 8.9GB | q8_0 | 4K | 11.1 GB |
| llama-pro:8b-instruct-fp16 | 17GB | fp16 | 4K context | 21.2 GB |
| llama-pro:8b-text-fp16 | 17GB | fp16 | 4K | 21.2 GB |
Strengths & Limitations
Strengths
- Programming expertise
- Mathematical reasoning
- General language understanding
Related models
deepscalerMath
A fine-tuned version of Deepseek-R1-Distilled-Qwen-1.5B that surpasses the performance of OpenAI’s o1-preview with just 1.5B parameters on popular math evaluations.
1.1M pullsqwen2-mathMath
Qwen2 Math is a series of specialized math language models built upon the Qwen2 LLMs, which significantly outperforms the mathematical capabilities of open-source models and even closed-source models (e.g., GPT4o).
450K pullswizard-mathMath
Model focused on math and logic problems
392K pullsmathstralMath
MathΣtral: a 7B model designed for math reasoning and scientific discovery by Mistral AI.
224K pulls