Mistral AI
Mixtral 8x7B Instruct v0.1
Check price at Amazon for Mistral AI Mixtral 8x7B Instruct v0.1 (affiliate link, opens in new tab)Affiliate link
- Weights at 16-bit
- 93 GB
- Weights at 8-bit (Q8_0)
- 50 GB
- Weights at 4-bit (Q4_K_M)
- 28 GB
- Smallest card at 4-bit
- GeForce RTX 5090 · 32 GB
- Parameters
- Active parameters
- Layers
- KV heads
- Context
- Licence
Specs
- Active parameters
- Attention heads
- Context
- Downloads (30 days)
- Experts
- Experts per token
- Family
- GGUF downloads
-
repo bytes files quant TheBloke/Mixtral-8x7B-Instruct-v0.1-GGUF 15,644,035,008 1 entries Q2_K TheBloke/Mixtral-8x7B-Instruct-v0.1-GGUF 20,363,356,096 1 entries Q3_K_M TheBloke/Mixtral-8x7B-Instruct-v0.1-GGUF 26,441,533,376 1 entries Q4_0 TheBloke/Mixtral-8x7B-Instruct-v0.1-GGUF 26,441,533,376 1 entries Q4_K_M TheBloke/Mixtral-8x7B-Instruct-v0.1-GGUF 32,229,279,680 1 entries Q5_0 TheBloke/Mixtral-8x7B-Instruct-v0.1-GGUF 32,229,279,680 1 entries Q5_K_M TheBloke/Mixtral-8x7B-Instruct-v0.1-GGUF 38,378,760,128 1 entries Q6_K TheBloke/Mixtral-8x7B-Instruct-v0.1-GGUF 49,624,262,592 1 entries Q8_0 - Head size
- Hidden size
- KV heads
- Layers
- Licence
- Parameters
- Parameters by type
- Publisher
- Released
More from Mistral AI
Across the range 4 models by family
MixtralMistral AI Mixtral 8x7B Instruct v0.1 · 32,768 tokens