diff --git a/comfy/latent_formats.py b/comfy/latent_formats.py
index 4b4a9eda..34c7bb3d 100644
--- a/comfy/latent_formats.py
+++ b/comfy/latent_formats.py
@@ -139,3 +139,14 @@ class SD3(LatentFormat):
 
 class StableAudio1(LatentFormat):
     latent_channels = 64
+
+class Flux(SD3):
+    def __init__(self):
+        self.scale_factor = 0.3611
+        self.shift_factor = 0.1159
+
+    def process_in(self, latent):
+        return (latent - self.shift_factor) * self.scale_factor
+
+    def process_out(self, latent):
+        return (latent / self.scale_factor) + self.shift_factor
diff --git a/comfy/ldm/flux/layers.py b/comfy/ldm/flux/layers.py
new file mode 100644
index 00000000..bb5e02b6
--- /dev/null
+++ b/comfy/ldm/flux/layers.py
@@ -0,0 +1,257 @@
+import math
+from dataclasses import dataclass
+
+import torch
+from einops import rearrange
+from torch import Tensor, nn
+
+from .math import attention, rope
+import comfy.ops
+
+
+class EmbedND(nn.Module):
+    def __init__(self, dim: int, theta: int, axes_dim: list[int]):
+        super().__init__()
+        self.dim = dim
+        self.theta = theta
+        self.axes_dim = axes_dim
+
+    def forward(self, ids: Tensor) -> Tensor:
+        n_axes = ids.shape[-1]
+        emb = torch.cat(
+            [rope(ids[..., i], self.axes_dim[i], self.theta) for i in range(n_axes)],
+            dim=-3,
+        )
+
+        return emb.unsqueeze(1)
+
+
+def timestep_embedding(t: Tensor, dim, max_period=10000, time_factor: float = 1000.0):
+    """
+    Create sinusoidal timestep embeddings.
+    :param t: a 1-D Tensor of N indices, one per batch element.
+                      These may be fractional.
+    :param dim: the dimension of the output.
+    :param max_period: controls the minimum frequency of the embeddings.
+    :return: an (N, D) Tensor of positional embeddings.
+    """
+    t = time_factor * t
+    half = dim // 2
+    freqs = torch.exp(-math.log(max_period) * torch.arange(start=0, end=half, dtype=torch.float32) / half).to(
+        t.device
+    )
+
+    args = t[:, None].float() * freqs[None]
+    embedding = torch.cat([torch.cos(args), torch.sin(args)], dim=-1)
+    if dim % 2:
+        embedding = torch.cat([embedding, torch.zeros_like(embedding[:, :1])], dim=-1)
+    if torch.is_floating_point(t):
+        embedding = embedding.to(t)
+    return embedding
+
+
+class MLPEmbedder(nn.Module):
+    def __init__(self, in_dim: int, hidden_dim: int, dtype=None, device=None, operations=None):
+        super().__init__()
+        self.in_layer = operations.Linear(in_dim, hidden_dim, bias=True, dtype=dtype, device=device)
+        self.silu = nn.SiLU()
+        self.out_layer = operations.Linear(hidden_dim, hidden_dim, bias=True, dtype=dtype, device=device)
+
+    def forward(self, x: Tensor) -> Tensor:
+        return self.out_layer(self.silu(self.in_layer(x)))
+
+
+class RMSNorm(torch.nn.Module):
+    def __init__(self, dim: int, dtype=None, device=None, operations=None):
+        super().__init__()
+        self.scale = nn.Parameter(torch.empty((dim), dtype=dtype, device=device))
+
+    def forward(self, x: Tensor):
+        x_dtype = x.dtype
+        x = x.float()
+        rrms = torch.rsqrt(torch.mean(x**2, dim=-1, keepdim=True) + 1e-6)
+        return (x * rrms).to(dtype=x_dtype) * comfy.ops.cast_to(self.scale, dtype=x_dtype, device=x.device)
+
+
+class QKNorm(torch.nn.Module):
+    def __init__(self, dim: int, dtype=None, device=None, operations=None):
+        super().__init__()
+        self.query_norm = RMSNorm(dim, dtype=dtype, device=device, operations=operations)
+        self.key_norm = RMSNorm(dim, dtype=dtype, device=device, operations=operations)
+
+    def forward(self, q: Tensor, k: Tensor, v: Tensor) -> tuple[Tensor, Tensor]:
+        q = self.query_norm(q)
+        k = self.key_norm(k)
+        return q.to(v), k.to(v)
+
+
+class SelfAttention(nn.Module):
+    def __init__(self, dim: int, num_heads: int = 8, qkv_bias: bool = False, dtype=None, device=None, operations=None):
+        super().__init__()
+        self.num_heads = num_heads
+        head_dim = dim // num_heads
+
+        self.qkv = operations.Linear(dim, dim * 3, bias=qkv_bias, dtype=dtype, device=device)
+        self.norm = QKNorm(head_dim, dtype=dtype, device=device, operations=operations)
+        self.proj = operations.Linear(dim, dim, dtype=dtype, device=device)
+
+    def forward(self, x: Tensor, pe: Tensor) -> Tensor:
+        qkv = self.qkv(x)
+        q, k, v = rearrange(qkv, "B L (K H D) -> K B H L D", K=3, H=self.num_heads)
+        q, k = self.norm(q, k, v)
+        x = attention(q, k, v, pe=pe)
+        x = self.proj(x)
+        return x
+
+
+@dataclass
+class ModulationOut:
+    shift: Tensor
+    scale: Tensor
+    gate: Tensor
+
+
+class Modulation(nn.Module):
+    def __init__(self, dim: int, double: bool, dtype=None, device=None, operations=None):
+        super().__init__()
+        self.is_double = double
+        self.multiplier = 6 if double else 3
+        self.lin = operations.Linear(dim, self.multiplier * dim, bias=True, dtype=dtype, device=device)
+
+    def forward(self, vec: Tensor) -> tuple[ModulationOut, ModulationOut | None]:
+        out = self.lin(nn.functional.silu(vec))[:, None, :].chunk(self.multiplier, dim=-1)
+
+        return (
+            ModulationOut(*out[:3]),
+            ModulationOut(*out[3:]) if self.is_double else None,
+        )
+
+
+class DoubleStreamBlock(nn.Module):
+    def __init__(self, hidden_size: int, num_heads: int, mlp_ratio: float, qkv_bias: bool = False, dtype=None, device=None, operations=None):
+        super().__init__()
+
+        mlp_hidden_dim = int(hidden_size * mlp_ratio)
+        self.num_heads = num_heads
+        self.hidden_size = hidden_size
+        self.img_mod = Modulation(hidden_size, double=True, dtype=dtype, device=device, operations=operations)
+        self.img_norm1 = operations.LayerNorm(hidden_size, elementwise_affine=False, eps=1e-6, dtype=dtype, device=device)
+        self.img_attn = SelfAttention(dim=hidden_size, num_heads=num_heads, qkv_bias=qkv_bias, dtype=dtype, device=device, operations=operations)
+
+        self.img_norm2 = operations.LayerNorm(hidden_size, elementwise_affine=False, eps=1e-6, dtype=dtype, device=device)
+        self.img_mlp = nn.Sequential(
+            operations.Linear(hidden_size, mlp_hidden_dim, bias=True, dtype=dtype, device=device),
+            nn.GELU(approximate="tanh"),
+            operations.Linear(mlp_hidden_dim, hidden_size, bias=True, dtype=dtype, device=device),
+        )
+
+        self.txt_mod = Modulation(hidden_size, double=True, dtype=dtype, device=device, operations=operations)
+        self.txt_norm1 = operations.LayerNorm(hidden_size, elementwise_affine=False, eps=1e-6, dtype=dtype, device=device)
+        self.txt_attn = SelfAttention(dim=hidden_size, num_heads=num_heads, qkv_bias=qkv_bias, dtype=dtype, device=device, operations=operations)
+
+        self.txt_norm2 = operations.LayerNorm(hidden_size, elementwise_affine=False, eps=1e-6, dtype=dtype, device=device)
+        self.txt_mlp = nn.Sequential(
+            operations.Linear(hidden_size, mlp_hidden_dim, bias=True, dtype=dtype, device=device),
+            nn.GELU(approximate="tanh"),
+            operations.Linear(mlp_hidden_dim, hidden_size, bias=True, dtype=dtype, device=device),
+        )
+
+    def forward(self, img: Tensor, txt: Tensor, vec: Tensor, pe: Tensor) -> tuple[Tensor, Tensor]:
+        img_mod1, img_mod2 = self.img_mod(vec)
+        txt_mod1, txt_mod2 = self.txt_mod(vec)
+
+        # prepare image for attention
+        img_modulated = self.img_norm1(img)
+        img_modulated = (1 + img_mod1.scale) * img_modulated + img_mod1.shift
+        img_qkv = self.img_attn.qkv(img_modulated)
+        img_q, img_k, img_v = rearrange(img_qkv, "B L (K H D) -> K B H L D", K=3, H=self.num_heads)
+        img_q, img_k = self.img_attn.norm(img_q, img_k, img_v)
+
+        # prepare txt for attention
+        txt_modulated = self.txt_norm1(txt)
+        txt_modulated = (1 + txt_mod1.scale) * txt_modulated + txt_mod1.shift
+        txt_qkv = self.txt_attn.qkv(txt_modulated)
+        txt_q, txt_k, txt_v = rearrange(txt_qkv, "B L (K H D) -> K B H L D", K=3, H=self.num_heads)
+        txt_q, txt_k = self.txt_attn.norm(txt_q, txt_k, txt_v)
+
+        # run actual attention
+        q = torch.cat((txt_q, img_q), dim=2)
+        k = torch.cat((txt_k, img_k), dim=2)
+        v = torch.cat((txt_v, img_v), dim=2)
+
+        attn = attention(q, k, v, pe=pe)
+        txt_attn, img_attn = attn[:, : txt.shape[1]], attn[:, txt.shape[1] :]
+
+        # calculate the img bloks
+        img = img + img_mod1.gate * self.img_attn.proj(img_attn)
+        img = img + img_mod2.gate * self.img_mlp((1 + img_mod2.scale) * self.img_norm2(img) + img_mod2.shift)
+
+        # calculate the txt bloks
+        txt = txt + txt_mod1.gate * self.txt_attn.proj(txt_attn)
+        txt = txt + txt_mod2.gate * self.txt_mlp((1 + txt_mod2.scale) * self.txt_norm2(txt) + txt_mod2.shift)
+        return img, txt
+
+
+class SingleStreamBlock(nn.Module):
+    """
+    A DiT block with parallel linear layers as described in
+    https://arxiv.org/abs/2302.05442 and adapted modulation interface.
+    """
+
+    def __init__(
+        self,
+        hidden_size: int,
+        num_heads: int,
+        mlp_ratio: float = 4.0,
+        qk_scale: float | None = None,
+        dtype=None,
+        device=None,
+        operations=None
+    ):
+        super().__init__()
+        self.hidden_dim = hidden_size
+        self.num_heads = num_heads
+        head_dim = hidden_size // num_heads
+        self.scale = qk_scale or head_dim**-0.5
+
+        self.mlp_hidden_dim = int(hidden_size * mlp_ratio)
+        # qkv and mlp_in
+        self.linear1 = operations.Linear(hidden_size, hidden_size * 3 + self.mlp_hidden_dim, dtype=dtype, device=device)
+        # proj and mlp_out
+        self.linear2 = operations.Linear(hidden_size + self.mlp_hidden_dim, hidden_size, dtype=dtype, device=device)
+
+        self.norm = QKNorm(head_dim, dtype=dtype, device=device, operations=operations)
+
+        self.hidden_size = hidden_size
+        self.pre_norm = operations.LayerNorm(hidden_size, elementwise_affine=False, eps=1e-6, dtype=dtype, device=device)
+
+        self.mlp_act = nn.GELU(approximate="tanh")
+        self.modulation = Modulation(hidden_size, double=False, dtype=dtype, device=device, operations=operations)
+
+    def forward(self, x: Tensor, vec: Tensor, pe: Tensor) -> Tensor:
+        mod, _ = self.modulation(vec)
+        x_mod = (1 + mod.scale) * self.pre_norm(x) + mod.shift
+        qkv, mlp = torch.split(self.linear1(x_mod), [3 * self.hidden_size, self.mlp_hidden_dim], dim=-1)
+
+        q, k, v = rearrange(qkv, "B L (K H D) -> K B H L D", K=3, H=self.num_heads)
+        q, k = self.norm(q, k, v)
+
+        # compute attention
+        attn = attention(q, k, v, pe=pe)
+        # compute activation in mlp stream, cat again and run second linear layer
+        output = self.linear2(torch.cat((attn, self.mlp_act(mlp)), 2))
+        return x + mod.gate * output
+
+
+class LastLayer(nn.Module):
+    def __init__(self, hidden_size: int, patch_size: int, out_channels: int, dtype=None, device=None, operations=None):
+        super().__init__()
+        self.norm_final = operations.LayerNorm(hidden_size, elementwise_affine=False, eps=1e-6, dtype=dtype, device=device)
+        self.linear = operations.Linear(hidden_size, patch_size * patch_size * out_channels, bias=True, dtype=dtype, device=device)
+        self.adaLN_modulation = nn.Sequential(nn.SiLU(), operations.Linear(hidden_size, 2 * hidden_size, bias=True, dtype=dtype, device=device))
+
+    def forward(self, x: Tensor, vec: Tensor) -> Tensor:
+        shift, scale = self.adaLN_modulation(vec).chunk(2, dim=1)
+        x = (1 + scale[:, None, :]) * self.norm_final(x) + shift[:, None, :]
+        x = self.linear(x)
+        return x
diff --git a/comfy/ldm/flux/math.py b/comfy/ldm/flux/math.py
new file mode 100644
index 00000000..e4ef624e
--- /dev/null
+++ b/comfy/ldm/flux/math.py
@@ -0,0 +1,29 @@
+import torch
+from einops import rearrange
+from torch import Tensor
+from comfy.ldm.modules.attention import optimized_attention
+
+def attention(q: Tensor, k: Tensor, v: Tensor, pe: Tensor) -> Tensor:
+    q, k = apply_rope(q, k, pe)
+
+    heads = q.shape[1]
+    x = optimized_attention(q, k, v, heads, skip_reshape=True)
+    return x
+
+
+def rope(pos: Tensor, dim: int, theta: int) -> Tensor:
+    assert dim % 2 == 0
+    scale = torch.arange(0, dim, 2, dtype=torch.float64, device=pos.device) / dim
+    omega = 1.0 / (theta**scale)
+    out = torch.einsum("...n,d->...nd", pos, omega)
+    out = torch.stack([torch.cos(out), -torch.sin(out), torch.sin(out), torch.cos(out)], dim=-1)
+    out = rearrange(out, "b n d (i j) -> b n d i j", i=2, j=2)
+    return out.float()
+
+
+def apply_rope(xq: Tensor, xk: Tensor, freqs_cis: Tensor) -> tuple[Tensor, Tensor]:
+    xq_ = xq.float().reshape(*xq.shape[:-1], -1, 1, 2)
+    xk_ = xk.float().reshape(*xk.shape[:-1], -1, 1, 2)
+    xq_out = freqs_cis[..., 0] * xq_[..., 0] + freqs_cis[..., 1] * xq_[..., 1]
+    xk_out = freqs_cis[..., 0] * xk_[..., 0] + freqs_cis[..., 1] * xk_[..., 1]
+    return xq_out.reshape(*xq.shape).type_as(xq), xk_out.reshape(*xk.shape).type_as(xk)
diff --git a/comfy/ldm/flux/model.py b/comfy/ldm/flux/model.py
new file mode 100644
index 00000000..f77834c1
--- /dev/null
+++ b/comfy/ldm/flux/model.py
@@ -0,0 +1,136 @@
+#Original code can be found on: https://github.com/black-forest-labs/flux
+
+from dataclasses import dataclass
+
+import torch
+from torch import Tensor, nn
+
+from .layers import (
+    DoubleStreamBlock,
+    EmbedND,
+    LastLayer,
+    MLPEmbedder,
+    SingleStreamBlock,
+    timestep_embedding,
+)
+
+from einops import rearrange, repeat
+
+@dataclass
+class FluxParams:
+    in_channels: int
+    vec_in_dim: int
+    context_in_dim: int
+    hidden_size: int
+    mlp_ratio: float
+    num_heads: int
+    depth: int
+    depth_single_blocks: int
+    axes_dim: list[int]
+    theta: int
+    qkv_bias: bool
+    guidance_embed: bool
+
+
+class Flux(nn.Module):
+    """
+    Transformer model for flow matching on sequences.
+    """
+
+    def __init__(self, image_model=None, dtype=None, device=None, operations=None, **kwargs):
+        super().__init__()
+        self.dtype = dtype
+        params = FluxParams(**kwargs)
+        self.params = params
+        self.in_channels = params.in_channels
+        self.out_channels = self.in_channels
+        if params.hidden_size % params.num_heads != 0:
+            raise ValueError(
+                f"Hidden size {params.hidden_size} must be divisible by num_heads {params.num_heads}"
+            )
+        pe_dim = params.hidden_size // params.num_heads
+        if sum(params.axes_dim) != pe_dim:
+            raise ValueError(f"Got {params.axes_dim} but expected positional dim {pe_dim}")
+        self.hidden_size = params.hidden_size
+        self.num_heads = params.num_heads
+        self.pe_embedder = EmbedND(dim=pe_dim, theta=params.theta, axes_dim=params.axes_dim)
+        self.img_in = operations.Linear(self.in_channels, self.hidden_size, bias=True, dtype=dtype, device=device)
+        self.time_in = MLPEmbedder(in_dim=256, hidden_dim=self.hidden_size, dtype=dtype, device=device, operations=operations)
+        self.vector_in = MLPEmbedder(params.vec_in_dim, self.hidden_size, dtype=dtype, device=device, operations=operations)
+        self.guidance_in = (
+            MLPEmbedder(in_dim=256, hidden_dim=self.hidden_size, dtype=dtype, device=device, operations=operations) if params.guidance_embed else nn.Identity()
+        )
+        self.txt_in = operations.Linear(params.context_in_dim, self.hidden_size, dtype=dtype, device=device)
+
+        self.double_blocks = nn.ModuleList(
+            [
+                DoubleStreamBlock(
+                    self.hidden_size,
+                    self.num_heads,
+                    mlp_ratio=params.mlp_ratio,
+                    qkv_bias=params.qkv_bias,
+                    dtype=dtype, device=device, operations=operations
+                )
+                for _ in range(params.depth)
+            ]
+        )
+
+        self.single_blocks = nn.ModuleList(
+            [
+                SingleStreamBlock(self.hidden_size, self.num_heads, mlp_ratio=params.mlp_ratio, dtype=dtype, device=device, operations=operations)
+                for _ in range(params.depth_single_blocks)
+            ]
+        )
+
+        self.final_layer = LastLayer(self.hidden_size, 1, self.out_channels, dtype=dtype, device=device, operations=operations)
+
+    def forward_orig(
+        self,
+        img: Tensor,
+        img_ids: Tensor,
+        txt: Tensor,
+        txt_ids: Tensor,
+        timesteps: Tensor,
+        y: Tensor,
+        guidance: Tensor | None = None,
+    ) -> Tensor:
+        if img.ndim != 3 or txt.ndim != 3:
+            raise ValueError("Input img and txt tensors must have 3 dimensions.")
+
+        # running on sequences img
+        img = self.img_in(img)
+        vec = self.time_in(timestep_embedding(timesteps, 256).to(img.dtype))
+        if self.params.guidance_embed:
+            if guidance is None:
+                raise ValueError("Didn't get guidance strength for guidance distilled model.")
+            vec = vec + self.guidance_in(timestep_embedding(guidance, 256).to(img.dtype))
+
+        vec = vec + self.vector_in(y)
+        txt = self.txt_in(txt)
+
+        ids = torch.cat((txt_ids, img_ids), dim=1)
+        pe = self.pe_embedder(ids)
+
+        for block in self.double_blocks:
+            img, txt = block(img=img, txt=txt, vec=vec, pe=pe)
+
+        img = torch.cat((txt, img), 1)
+        for block in self.single_blocks:
+            img = block(img, vec=vec, pe=pe)
+        img = img[:, txt.shape[1] :, ...]
+
+        img = self.final_layer(img, vec)  # (N, T, patch_size ** 2 * out_channels)
+        return img
+
+    def forward(self, x, timestep, context, y, guidance, **kwargs):
+        bs, c, h, w = x.shape
+        img = rearrange(x, "b c (h ph) (w pw) -> b (h w) (c ph pw)", ph=2, pw=2)
+
+        img_ids = torch.zeros((h // 2, w // 2, 3), device=x.device, dtype=x.dtype)
+        img_ids[..., 1] = img_ids[..., 1] + torch.arange(h // 2, device=x.device, dtype=x.dtype)[:, None]
+        img_ids[..., 2] = img_ids[..., 2] + torch.arange(w // 2, device=x.device, dtype=x.dtype)[None, :]
+        img_ids = repeat(img_ids, "h w c -> b (h w) c", b=bs)
+
+        txt_ids = torch.zeros((bs, context.shape[1], 3), device=x.device, dtype=x.dtype)
+        out = self.forward_orig(img, img_ids, context, txt_ids, timestep, y, guidance)
+        return rearrange(out, "b (h w) (c ph pw) -> b c (h ph) (w pw)", h=h // 2, w=w // 2, ph=2, pw=2)
diff --git a/comfy/model_base.py b/comfy/model_base.py
index 9d60c1fb..7c7b4c3f 100644
--- a/comfy/model_base.py
+++ b/comfy/model_base.py
@@ -10,6 +10,8 @@ import comfy.ldm.aura.mmdit
 import comfy.ldm.hydit.models
 import comfy.ldm.audio.dit
 import comfy.ldm.audio.embedders
+import comfy.ldm.flux.model
+
 import comfy.model_management
 import comfy.conds
 import comfy.ops
@@ -26,6 +28,7 @@ class ModelType(Enum):
     EDM = 5
     FLOW = 6
     V_PREDICTION_CONTINUOUS = 7
+    FLUX = 8
 
 
 from comfy.model_sampling import EPS, V_PREDICTION, EDM, ModelSamplingDiscrete, ModelSamplingContinuousEDM, StableCascadeSampling, ModelSamplingContinuousV
@@ -53,6 +56,9 @@ def model_sampling(model_config, model_type):
     elif model_type == ModelType.V_PREDICTION_CONTINUOUS:
         c = V_PREDICTION
         s = ModelSamplingContinuousV
+    elif model_type == ModelType.FLUX:
+        c = comfy.model_sampling.CONST
+        s = comfy.model_sampling.ModelSamplingFlux
 
     class ModelSampling(s, c):
         pass
@@ -681,3 +687,18 @@ class HunyuanDiT(BaseModel):
 
         out['image_meta_size'] = comfy.conds.CONDRegular(torch.FloatTensor([[height, width, target_height, target_width, 0, 0]]))
         return out
+
+class Flux(BaseModel):
+    def __init__(self, model_config, model_type=ModelType.FLUX, device=None):
+        super().__init__(model_config, model_type, device=device, unet_model=comfy.ldm.flux.model.Flux)
+
+    def encode_adm(self, **kwargs):
+        return kwargs["pooled_output"]
+
+    def extra_conds(self, **kwargs):
+        out = super().extra_conds(**kwargs)
+        cross_attn = kwargs.get("cross_attn", None)
+        if cross_attn is not None:
+            out['c_crossattn'] = comfy.conds.CONDRegular(cross_attn)
+        out['guidance'] = comfy.conds.CONDRegular(torch.FloatTensor([kwargs.get("guidance", 3.5)]))
+        return out
diff --git a/comfy/model_detection.py b/comfy/model_detection.py
index ea495937..dda9797b 100644
--- a/comfy/model_detection.py
+++ b/comfy/model_detection.py
@@ -128,6 +128,23 @@ def detect_unet_config(state_dict, key_prefix):
             unet_config["image_model"] = "hydit1"
         return unet_config
 
+    if '{}double_blocks.0.img_attn.norm.key_norm.scale'.format(key_prefix) in state_dict_keys: #Flux
+        dit_config = {}
+        dit_config["image_model"] = "flux"
+        dit_config["in_channels"] = 64
+        dit_config["vec_in_dim"] = 768
+        dit_config["context_in_dim"] = 4096
+        dit_config["hidden_size"] = 3072
+        dit_config["mlp_ratio"] = 4.0
+        dit_config["num_heads"] = 24
+        dit_config["depth"] = 19
+        dit_config["depth_single_blocks"] = 38
+        dit_config["axes_dim"] = [16, 56, 56]
+        dit_config["theta"] = 10000
+        dit_config["qkv_bias"] = True
+        dit_config["guidance_embed"] = "{}guidance_in.in_layer.weight".format(key_prefix) in state_dict_keys
+        return dit_config
+
     if '{}input_blocks.0.0.weight'.format(key_prefix) not in state_dict_keys:
         return None
 
diff --git a/comfy/model_sampling.py b/comfy/model_sampling.py
index 25bb7e04..4a0f2db6 100644
--- a/comfy/model_sampling.py
+++ b/comfy/model_sampling.py
@@ -272,3 +272,43 @@ class StableCascadeSampling(ModelSamplingDiscrete):
 
         percent = 1.0 - percent
         return self.sigma(torch.tensor(percent))
+
+
+def flux_time_shift(mu: float, sigma: float, t):
+    return math.exp(mu) / (math.exp(mu) + (1 / t - 1) ** sigma)
+
+class ModelSamplingFlux(torch.nn.Module):
+    def __init__(self, model_config=None):
+        super().__init__()
+        if model_config is not None:
+            sampling_settings = model_config.sampling_settings
+        else:
+            sampling_settings = {}
+
+        self.set_parameters(shift=sampling_settings.get("shift", 1.15))
+
+    def set_parameters(self, shift=1.15, timesteps=10000):
+        self.shift = shift
+        ts = self.sigma((torch.arange(1, timesteps + 1, 1) / timesteps))
+        self.register_buffer('sigmas', ts)
+
+    @property
+    def sigma_min(self):
+        return self.sigmas[0]
+
+    @property
+    def sigma_max(self):
+        return self.sigmas[-1]
+
+    def timestep(self, sigma):
+        return sigma
+
+    def sigma(self, timestep):
+        return flux_time_shift(self.shift, 1.0, timestep)
+
+    def percent_to_sigma(self, percent):
+        if percent <= 0.0:
+            return 1.0
+        if percent >= 1.0:
+            return 0.0
+        return 1.0 - percent
diff --git a/comfy/sd.py b/comfy/sd.py
index 8bf8d108..c9bc1639 100644
--- a/comfy/sd.py
+++ b/comfy/sd.py
@@ -23,6 +23,7 @@ import comfy.text_encoders.sd3_clip
 import comfy.text_encoders.sa_t5
 import comfy.text_encoders.aura_t5
 import comfy.text_encoders.hydit
+import comfy.text_encoders.flux
 
 import comfy.model_patcher
 import comfy.lora
@@ -387,6 +388,7 @@ class CLIPType(Enum):
     SD3 = 3
     STABLE_AUDIO = 4
     HUNYUAN_DIT = 5
+    FLUX = 6
 
 def load_clip(ckpt_paths, embedding_directory=None, clip_type=CLIPType.STABLE_DIFFUSION):
     clip_data = []
@@ -438,6 +440,9 @@ def load_clip(ckpt_paths, embedding_directory=None, clip_type=CLIPType.STABLE_DI
         elif clip_type == CLIPType.HUNYUAN_DIT:
             clip_target.clip = comfy.text_encoders.hydit.HyditModel
             clip_target.tokenizer = comfy.text_encoders.hydit.HyditTokenizer
+        elif clip_type == CLIPType.FLUX:
+            clip_target.clip = comfy.text_encoders.flux.FluxClipModel
+            clip_target.tokenizer = comfy.text_encoders.flux.FluxTokenizer
         else:
             clip_target.clip = sdxl_clip.SDXLClipModel
             clip_target.tokenizer = sdxl_clip.SDXLTokenizer
diff --git a/comfy/supported_models.py b/comfy/supported_models.py
index ddd0c173..43e8f5d1 100644
--- a/comfy/supported_models.py
+++ b/comfy/supported_models.py
@@ -9,6 +9,7 @@ import comfy.text_encoders.sd3_clip
 import comfy.text_encoders.sa_t5
 import comfy.text_encoders.aura_t5
 import comfy.text_encoders.hydit
+import comfy.text_encoders.flux
 
 from . import supported_models_base
 from . import latent_formats
@@ -619,7 +620,45 @@ class HunyuanDiT1(HunyuanDiT):
         "linear_end" : 0.03,
     }
 
+class Flux(supported_models_base.BASE):
+    unet_config = {
+        "image_model": "flux",
+        "guidance_embed": True,
+    }
 
-models = [Stable_Zero123, SD15_instructpix2pix, SD15, SD20, SD21UnclipL, SD21UnclipH, SDXL_instructpix2pix, SDXLRefiner, SDXL, SSD1B, KOALA_700M, KOALA_1B, Segmind_Vega, SD_X4Upscaler, Stable_Cascade_C, Stable_Cascade_B, SV3D_u, SV3D_p, SD3, StableAudio, AuraFlow, HunyuanDiT, HunyuanDiT1]
+    sampling_settings = {
+    }
+
+    unet_extra_config = {}
+    latent_format = latent_formats.Flux
+    supported_inference_dtypes = [torch.bfloat16, torch.float32]
+
+    vae_key_prefix = ["vae."]
+    text_encoder_key_prefix = ["text_encoders."]
+
+    def get_model(self, state_dict, prefix="", device=None):
+        out = model_base.Flux(self, device=device)
+        return out
+
+    def clip_target(self, state_dict={}):
+        return supported_models_base.ClipTarget(comfy.text_encoders.flux.FluxTokenizer, comfy.text_encoders.flux.FluxClipModel)
+
+class FluxSchnell(Flux):
+    unet_config = {
+        "image_model": "flux",
+        "guidance_embed": False,
+    }
+
+    sampling_settings = {
+        "multiplier": 1.0,
+        "shift": 1.0,
+    }
+
+    def get_model(self, state_dict, prefix="", device=None):
+        out = model_base.Flux(self, model_type=model_base.ModelType.FLOW, device=device)
+        return out
+
+
+models = [Stable_Zero123, SD15_instructpix2pix, SD15, SD20, SD21UnclipL, SD21UnclipH, SDXL_instructpix2pix, SDXLRefiner, SDXL, SSD1B, KOALA_700M, KOALA_1B, Segmind_Vega, SD_X4Upscaler, Stable_Cascade_C, Stable_Cascade_B, SV3D_u, SV3D_p, SD3, StableAudio, AuraFlow, HunyuanDiT, HunyuanDiT1, Flux, FluxSchnell]
 
 models += [SVD_img2vid]
diff --git a/comfy/text_encoders/flux.py b/comfy/text_encoders/flux.py
new file mode 100644
index 00000000..2759a38a
--- /dev/null
+++ b/comfy/text_encoders/flux.py
@@ -0,0 +1,64 @@
+from comfy import sd1_clip
+import comfy.text_encoders.t5
+from transformers import T5TokenizerFast
+import torch
+import os
+
+class T5XXLModel(sd1_clip.SDClipModel):
+    def __init__(self, device="cpu", layer="last", layer_idx=None, dtype=None):
+        textmodel_json_config = os.path.join(os.path.dirname(os.path.realpath(__file__)), "t5_config_xxl.json")
+        super().__init__(device=device, layer=layer, layer_idx=layer_idx, textmodel_json_config=textmodel_json_config, dtype=dtype, special_tokens={"end": 1, "pad": 0}, model_class=comfy.text_encoders.t5.T5)
+
+class T5XXLTokenizer(sd1_clip.SDTokenizer):
+    def __init__(self, embedding_directory=None, tokenizer_data={}):
+        tokenizer_path = os.path.join(os.path.dirname(os.path.realpath(__file__)), "t5_tokenizer")
+        super().__init__(tokenizer_path, pad_with_end=False, embedding_size=4096, embedding_key='t5xxl', tokenizer_class=T5TokenizerFast, has_start_token=False, pad_to_max_length=False, max_length=99999999, min_length=256)
+
+
+class FluxTokenizer:
+    def __init__(self, embedding_directory=None, tokenizer_data={}):
+        self.clip_l = sd1_clip.SDTokenizer(embedding_directory=embedding_directory)
+        self.t5xxl = T5XXLTokenizer(embedding_directory=embedding_directory)
+
+    def tokenize_with_weights(self, text:str, return_word_ids=False):
+        out = {}
+        out["l"] = self.clip_l.tokenize_with_weights(text, return_word_ids)
+        out["t5xxl"] = self.t5xxl.tokenize_with_weights(text, return_word_ids)
+        return out
+
+    def untokenize(self, token_weight_pair):
+        return self.clip_g.untokenize(token_weight_pair)
+
+    def state_dict(self):
+        return {}
+
+
+class FluxClipModel(torch.nn.Module):
+    def __init__(self, device="cpu", dtype=None):
+        super().__init__()
+        self.clip_l = sd1_clip.SDClipModel(device=device, dtype=dtype, return_projected_pooled=False)
+        self.t5xxl = T5XXLModel(device=device, dtype=dtype)
+        self.dtypes = set([dtype])
+
+    def set_clip_options(self, options):
+        self.clip_l.set_clip_options(options)
+        self.t5xxl.set_clip_options(options)
+
+    def reset_clip_options(self):
+        self.clip_l.reset_clip_options()
+        self.t5xxl.reset_clip_options()
+
+    def encode_token_weights(self, token_weight_pairs):
+        token_weight_pairs_l = token_weight_pairs["l"]
+        token_weight_pars_t5 = token_weight_pairs["t5xxl"]
+
+        t5_out, t5_pooled = self.t5xxl.encode_token_weights(token_weight_pars_t5)
+        l_out, l_pooled = self.clip_l.encode_token_weights(token_weight_pairs_l)
+        return t5_out, l_pooled
+
+    def load_sd(self, sd):
+        if "text_model.encoder.layers.1.mlp.fc1.weight" in sd:
+            return self.clip_l.load_sd(sd)
+        else:
+            return self.t5xxl.load_sd(sd)
+
diff --git a/folder_paths.py b/folder_paths.py
index 2baf8ce1..71faa2df 100644
--- a/folder_paths.py
+++ b/folder_paths.py
@@ -3,7 +3,7 @@ import time
 import logging
 from typing import Set, List, Dict, Tuple
 
-supported_pt_extensions: Set[str] = set(['.ckpt', '.pt', '.bin', '.pth', '.safetensors', '.pkl'])
+supported_pt_extensions: Set[str] = set(['.ckpt', '.pt', '.bin', '.pth', '.safetensors', '.pkl', '.sft'])
 
 SupportedFileExtensionsType = Set[str]
 ScanPathType = List[str]
diff --git a/nodes.py b/nodes.py
index dc66dd3e..93d24ae5 100644
--- a/nodes.py
+++ b/nodes.py
@@ -859,7 +859,7 @@ class DualCLIPLoader:
     def INPUT_TYPES(s):
         return {"required": { "clip_name1": (folder_paths.get_filename_list("clip"), ),
                               "clip_name2": (folder_paths.get_filename_list("clip"), ),
-                              "type": (["sdxl", "sd3"], ),
+                              "type": (["sdxl", "sd3", "flux"], ),
                              }}
     RETURN_TYPES = ("CLIP",)
     FUNCTION = "load_clip"
@@ -873,6 +873,8 @@ class DualCLIPLoader:
             clip_type = comfy.sd.CLIPType.STABLE_DIFFUSION
         elif type == "sd3":
             clip_type = comfy.sd.CLIPType.SD3
+        elif type == "flux":
+            clip_type = comfy.sd.CLIPType.FLUX
 
         clip = comfy.sd.load_clip(ckpt_paths=[clip_path1, clip_path2], embedding_directory=folder_paths.get_folder_paths("embeddings"), clip_type=clip_type)
         return (clip,)