Mistral · 2024-04 · Sparse MoE
Mixtral 8x22B
Mixtral 8x22B is a grouped-query attention (GQA) transformer released by Mistral in 2024-04, with 56 layers, hidden size 6144 and a context window of 65,536 tokens.
Layer stack
Attention ×56 · Layers 56
Key facts
| Family | Mistral |
|---|---|
| Released | 2024-04 |
| Params | 140.6 B |
| Active | 39.1 B |
| Context | 65,536 tokens |
| Attention | GQA (48:8) |
| Layers | 56 |
| Hidden | 6,144 |
| Heads | 48 |
| Vocab | 32,000 |
| Position | RoPE |
| Norm | RMSNorm |
| Activation | silu |
| Dtype | — |
| Architecture class | MixtralForCausalLM |
Architecture overview
Attention
grouped-query attention (GQA) — 48 q-heads / 8 kv-heads. RoPE θ=1,000,000.
Feed-forward / MoE
It is a sparse mixture-of-experts with 8 routed experts and top-2 routing; about 39.1B parameters are active per token.
Field-level comparison
Compared against the previous model of the same lab; where none exists, against the structurally closest model. The ratio column is this model divided by the comparison model.
Full-field comparison vs predecessor Mistral 7B
| Model | Mixtral 8x22B | Mistral 7B | Ratio |
|---|---|---|---|
| model_type | mixtral | mistral | ≠ |
| architectures | MixtralForCausalLM | MistralForCausalLM | ≠ |
| hidden_size | 6144 | 4096 | ×1.50 |
| num_hidden_layers | 56 | 32 | ×1.75 |
| num_attention_heads | 48 | 32 | ×1.50 |
| num_key_value_heads | 8 | 8 | ≈1 |
| intermediate_size | 16384 | 14336 | ×1.14 |
| hidden_act | silu | silu | ＝ |
| num_experts_per_tok | 2 | — | |
| rope_theta | 1000000 | 10000 | ×100.00 |
| max_position_embeddings | 65536 | 32768 | ×2.00 |
| sliding_window | — | 4096 | |
| rms_norm_eps | 0.00001 | 0.00001 | ≈1 |
| vocab_size | 32000 | 32000 | ≈1 |
| torch_dtype | bfloat16 | bfloat16 | ＝ |
| attention_dropout | 0 | — | |
| bos_token_id | 1 | 1 | ≈1 |
| eos_token_id | 2 | 2 | ≈1 |
| initializer_range | 0.02 | 0.02 | ≈1 |
| num_local_experts | 8 | — | |
| output_router_logits | false | — | |
| router_aux_loss_coef | 0.001 | — | |
| tie_word_embeddings | false | false | ＝ |
| transformers_version | 4.38.0 | 4.34.0.dev0 | ≠ |
| use_cache | true | true | ＝ |
Most similar architectures
- MiniMax-Text-01 (MiniMax, 2025-01) 0.988
- MiniMax-M1 (MiniMax, 2025-06) 0.988
- GLM-4.5-Air (GLM, 2025-07) 0.986
- GPT-OSS 20B (OpenAI, 2025-08) 0.984
- GPT-OSS 120B (OpenAI, 2025-08) 0.982
- Qwen3 235B A22B (Qwen, 2025-04) 0.981
Similarity values range from 0 (no shared categorical features) to 1 (identical profiles).
Raw config fields
24 fields
| architectures | MixtralForCausalLM |
|---|---|
| attention_dropout | 0 |
| bos_token_id | 1 |
| eos_token_id | 2 |
| hidden_act | silu |
| hidden_size | 6144 |
| initializer_range | 0.02 |
| intermediate_size | 16384 |
| max_position_embeddings | 65536 |
| model_type | mixtral |
| num_attention_heads | 48 |
| num_experts_per_tok | 2 |
| num_hidden_layers | 56 |
| num_key_value_heads | 8 |
| num_local_experts | 8 |
| output_router_logits | false |
| rms_norm_eps | 0.00001 |
| rope_theta | 1000000 |
| router_aux_loss_coef | 0.001 |
| tie_word_embeddings | false |
| torch_dtype | bfloat16 |
| transformers_version | 4.38.0 |
| use_cache | true |
| vocab_size | 32000 |