No machine on this network has measured this model yet. Everything below is what the catalog knows exists — where the weights come from, how large each build is, and whose work it is. Nothing here is a claim about what the model can do; that only appears once a machine has run it and signed the result.
unsloth/Kimi-K2-Instruct-GGUF:IQ1_Sllamacpp · IQ1_S · 260.9 GiB · quantizer (unsloth)
unsloth/Kimi-K2-Instruct-GGUF:IQ2_XXSllamacpp · IQ2_XXS · 306.2 GiB · quantizer (unsloth)
unsloth/Kimi-K2-Instruct-GGUF:IQ3_XXSllamacpp · IQ3_XXS · 388.0 GiB · quantizer (unsloth)
mlx-community/moonshotai_Kimi-K2-Instruct-mlx-3bitmlx · 3BIT · 418.5 GiB · quantizer (mlx-community)
mlx-community/Kimi-K2-Instruct-4bitmlx · 4BIT · 537.9 GiB · quantizer (mlx-community)
unsloth/Kimi-K2-Instruct-GGUF:Q4_0llamacpp · Q4_0 · 540.7 GiB · quantizer (unsloth)
unsloth/Kimi-K2-Instruct-GGUF:Q4_K_Mllamacpp · Q4_K_M · 578.1 GiB · quantizer (unsloth)
moonshotai/Kimi-K2-Instructvllm · FP8 · 958.5 GiB · vendor (moonshotai)
unsloth/Kimi-K2-Instruct-GGUF:Q8_0llamacpp · Q8_0 · 1016.1 GiB · quantizer (unsloth)
unsloth/Kimi-K2-Instruct-GGUF:BF16llamacpp · BF16 · 1912.1 GiB · quantizer (unsloth)