37 tok/s from Qwen3-Next-80B (48 GB) on a 32 GB laptop, 24 without drafts vs 2.8 for llama.cpp; reading 62 vs 16 tok/s. Faster than llama.cpp on all 13 models compared in Q4_K_M. LLMs bigger than your RAM, pure C, CPU or GPU.
C avx2ccpu-inferenceggufinference-enginelinuxllmlocal-llmmixture-of-expertsmoenvmequantizationqwen3simdspeculative-decodingvulkan
/* SPDX-License-Identifier: GPL-3.0-or-later
Copyright (C) 2026 Maurizio Cammalleri */
/*
* catalog.h - the models Janas knows: where each comes from, the SHA-256 of
* every file on the way, as MODELS.md lists them, and how janas-chat runs
* it fastest. janas-get downloads and converts them; janas-chat lists them,
* takes one by name and starts it with its profile.
*/
#ifndef JANAS_COMMON_CATALOG_H
#define JANAS_COMMON_CATALOG_H
#include <stddef.h>
#define JANAS_CATALOG_PARTS 4
/* how the model's multi-token prediction file is made, if it has one */
enum janas_model_extra {
JANAS_EXTRA_NONE,
JANAS_EXTRA_GGUF, /* a GGUF of its own (Gemma 4's assistant), gguf2jns */
JANAS_EXTRA_HF2JNS, /* the block of the original checkpoint, hf2jns_mtp
hf: */
JANAS_EXTRA_BLOCK /* a GGUF of the block alone (unsloth's of
Qwen3.8-Flash-Next), perhaps from another
repository, copied by hf2jns_mtp */
};
struct janas_model_entry {
const char *name; /* what janas-get takes */
const char *what; /* a line for the list */
const char *repo; /* the GGUF's repository */
const char *parts[JANAS_CATALOG_PARTS]; /* its file(s), the first given to
gguf2jns */
const char *part_sha[JANAS_CATALOG_PARTS];
double gguf_gb; /* download */
const char *out; /* the .jns, beside the downloads */
/* the same .jns (and the extra file), converted and checked, on
Hugging Face: what janas-get downloads unless --from-gguf; its
size with the extra file */
const char *jns_repo;
double jns_gb;
const char *flat_sha; /* what gguf2jns writes (the .jns itself when
there are no planes) */
const char *jns_sha; /* what jns_planes writes, NULL: one step */
enum janas_model_extra extra;
const char *extra_src; /* GGUF file of repo (of extra_repo, if
given), or hf:repo[@rev] */
const char *extra_repo;
const char *extra_src_sha; /* of that GGUF (JANAS_EXTRA_GGUF, _BLOCK) */
const char *extra_out;
const char *extra_sha;
int ram_gb; /* the smallest machine it is meant for */
int not_chat; /* 1: not a chat model (embeddings) */
int test_only; /* 1: for testing the engine, not for chatting: left
out of janas-chat's list, opened by name with a
word of warning */
const char *draft; /* the catalog model that guesses its tokens, when
that beats its own prediction file */
/* janas-chat's options for the "fast" profile, on top of the default
("quality") one: the levers measured to make this model faster at a
small, measured cost to its answers (README, ChangeLog) */
const char *fast;
};
extern const struct janas_model_entry janas_catalog[];
extern const int janas_catalog_n;
const struct janas_model_entry *janas_catalog_find(const char *name);
/*
* Where the models are: JANAS_MODELS if set, otherwise
* $XDG_DATA_HOME/janas/models, otherwise ~/.local/share/janas/models.
* Returns 0, or -1 if there is no home to put it in.
*/
int janas_models_dir(char *buf, size_t len);
/* Makes dir and its parents (mode 0755). 0 if it is there afterwards. */
int janas_mkdirs(const char *dir);
#endif