LLM ipi ni bora kwa Mac yenye GB 256?

Miongoni mwa modeli zenye Intelligence Index iliyochapishwa, Qwen3.8 Flash Next ina alama ya juu zaidi kati ya zile zinazoenda vizuri ndani ya GB 256 (6-bit). Kubwa zaidi inayoenda vizuri ni MiMo V2.6 Flash. Kwa code: Qwen3 Coder Next.

Je, GB 256 zinatosha kwa AI ya ndani kwenye Mac?

Modeli huria 54 kati ya 73 zilizo hapa zinaenda vizuri ndani yake kwa biti 4 au zaidi. Kwa chaguo-msingi macOS huiruhusu GPU kutumia takriban GB 192 kati yake; modeli na context yake lazima zitoshe humo.

Je, Mac ya GB 256 inaweza kuendesha gpt-oss-20b?

Ndiyo, 4.25-bit asili, kwa takriban tokeni 194–345 kwa sekunde kwenye M3 Ultra na 284–505 kwenye M5 Ultra.

LLM bora za ndani kwa Mac ya GB 256 (M3 hadi M5)

Mac zenye GB 256

Chipu 2 kuanzia M3 hadi M5 huja na GB 256. Fungua yoyote kuona orodha na kasi zake.

Zinaenda vizuri 61

Inatosha na kubakiza nafasi ya context, kwa tokeni 10 au zaidi kwa sekunde kwenye M5 Ultra.

Modeli	Toleo	Kumbukumbu	M3 Ultra	M5 Ultra
Qwen3.8 Flash Next180B · 6B zinazotumika · Akili 39.8	6-bit	146 GB	75–134	110–196
MiMo V2.6 Flash309B · 15B zinazotumika · Akili 37.9	4.48-bit asili	173 GB	44–78	64–114
Qwen3.8 27B27B · Akili 33.7	8-bit	28.7 GB	13–23	19–33
Inkling Small276B · 12B zinazotumika · Akili 25.7	4-bit	155 GB	54–96	79–141
Hy3295B · 21B zinazotumika · Akili 25.3	4-bit	166 GB	31–56	46–81
Nemotron 3.5 Bolt 74B A9B74B · 9B zinazotumika · Akili 21.4	8-bit	78.6 GB	38–68	56–100
Qwen3.6 35B A3B35B · 3B zinazotumika · Akili 18.2	8-bit	37.2 GB	115–205	169–300
Muse Glimmer 30B29.6B · Akili 17.5	8-bit	31.5 GB	12–21	17–30
Qwen3.5 122B A10B122B · 10B zinazotumika · Akili 15.6	8-bit	130 GB	35–62	51–91
Gemma 4 31B30.7B · Akili 14.7	8-bit	32.6 GB	11–20	17–29
Nemotron 3.5 Lightning 30B A3B30B · 3B zinazotumika · Akili 12.9	8-bit	31.9 GB	115–205	169–300
Granite 4.2 30B30B · Akili 12.7	8-bit	31.9 GB	12–21	17–30
MiniCPM5 2B2.52B · Akili 12.5	8-bit	2.7 GB	137–243	200–356
gpt-oss-120b117B · 5.1B zinazotumika · Akili 11.6	4.25-bit asili	65.3 GB	137–243	200–356
Mistral Small 4 (119B)119B · 6.5B zinazotumika · Akili 11.3	8-bit	126 GB	53–95	78–139
Qwen3.5 9B9B · Akili 11.2	8-bit	9.6 GB	38–68	56–100
Granite 4.2 8B8B · Akili 11.1	8-bit	8.5 GB	43–77	64–113
Qwen3 Coder Next80B · 3B zinazotumika · Akili 9.2	8-bit	85 GB	115–205	169–300
Granite 4.2 3B3B · Akili 9.1	8-bit	3.2 GB	115–205	169–300
gpt-oss-20b21B · 3.6B zinazotumika · Akili 9	4.25-bit asili	13.8 GB	194–345	284–505
Nemotron 3 Nano 30B A3B30B · 3.5B zinazotumika · Akili 8.9	8-bit	31.9 GB	100–177	146–259
Gemma 4 E4B8B · Akili 8.9	8-bit	8.5 GB	77–137	113–200
Llama 4 Scout 17B-16E Instruct109B · 17B zinazotumika · Akili 6.5	8-bit	116 GB	20–36	30–53
Ministral 3 8B8.9B · Akili 5.5	8-bit	9.5 GB	39–69	57–101
Nemotron 3 Nano 4B3.97B · Akili 4.8	8-bit	4.2 GB	88–156	129–229
DeepSeek V4 Flash 0731284B · 13B zinazotumika	4-bit	160 GB	50–90	74–132
MiniMax M2.7230B · 10B zinazotumika	4-bit	129 GB	66–117	96–171
Ornith 1.5 122B122B · 10B zinazotumika	8-bit	130 GB	35–62	51–91
Nemotron 3 Super 120B A12B120B · 12B zinazotumika	8-bit	128 GB	29–51	42–75
GLM-4.5 Air106B · 12B zinazotumika	8-bit	113 GB	29–51	42–75
Llama 3.3 70B Instruct70B	4-bit	39.4 GB	5–9	14–24
Qwen3 Coder 30B A3B Instruct30.5B · 3.3B zinazotumika	8-bit	32.4 GB	105–187	154–274
GLM-4.7 Flash30B · 3B zinazotumika	8-bit	31.9 GB	115–205	169–300
Qwen3.6 27B27B	8-bit	28.7 GB	13–23	19–33
Gemma 4 26B A4B25.2B · 3.8B zinazotumika	8-bit	26.8 GB	92–164	135–240
Devstral Small 2 (24B)24B	8-bit	25.5 GB	14–26	21–38
Magistral Small 1.2 (2509)24B	8-bit	25.5 GB	14–26	21–38
Mistral Small 3.2 (24B)24B	8-bit	25.5 GB	14–26	21–38
LFM2.5 24B A2B24B · 2B zinazotumika	8-bit	25.5 GB	176–312	257–457
Phi-4 Reasoning Vision 15B15B	8-bit	15.9 GB	23–41	34–60
Phi-4 (14B)14B	8-bit	14.9 GB	25–44	36–64
Gemma 4 12B Unified12B	8-bit	12.7 GB	29–52	43–76
Qwen3 8B8.2B	8-bit	8.7 GB	42–75	62–110
Llama 3.1 8B Instruct8B	8-bit	8.5 GB	43–77	64–113
Gemma 4 E2B5.1B	8-bit	5.4 GB	154–273	225–400
Qwen3.5 4B4B	8-bit	4.3 GB	86–152	126–223
Phi-4 mini instruct3.8B	8-bit	4 GB	92–164	135–240
Llama 3.2 3B Instruct3.21B	8-bit	3.4 GB	108–193	159–282
SmolLM3 3B3B	8-bit	3.2 GB	115–205	169–300
LFM2.5 2.6B2.69B	8-bit	2.9 GB	127–226	186–331
Qwen3.5 2B2B	8-bit	2.1 GB	176–312	257–457
Llama 3.2 1B Instruct1.23B	8-bit	1.3 GB	284–504	415–738
Qwen3.5 0.8B0.8B	8-bit	0.9 GB	410–728	600–1067
Qwen3 0.6B0.6B	8-bit	0.6 GB	614–1092	900–1600
DeepSeek V4.1 Flash 1017552B · 16B zinazotumika · Akili 45.1	2-bit ubora wa chini	173 GB	74–131	108–192
GLM-5.3 Flash320B · 18B zinazotumika · Akili 41.8	3-bit ubora wa chini	140 GB	47–83	68–122
Beam 1.0501B · 23B zinazotumika · Akili 30.3	2-bit ubora wa chini	157 GB	51–91	75–133
MiniMax M3428B · 23B zinazotumika · Akili 29.2	2-bit ubora wa chini	134 GB	51–91	75–133
Nemotron 3 Ultra 550B A55B550B · 55B zinazotumika · Akili 22.9	2-bit ubora wa chini	172 GB	21–38	31–56
Qwen3.5 397B A17B397B · 17B zinazotumika · Akili 18.4	3-bit ubora wa chini	174 GB	50–89	73–130
Llama 4 Maverick 17B-128E Instruct400B · 17B zinazotumika · Akili 9.3	3-bit ubora wa chini	175 GB	50–89	73–130

Zinaenda, lakini polepole 2

Inatosha, lakini chini ya tokeni 10 kwa sekunde: sawa kwa kazi za kundi, polepole kwa mazungumzo.

Modeli	Toleo	Kumbukumbu	M3 Ultra	M5 Ultra
Mistral Medium 3.5 (128B)128B · Akili 14.2	8-bit	136 GB	3–5	4–7
Devstral 2 (123B)123B	8-bit	131 GB	3–5	4–7

Zinatosha kwa shida 1

Inahitaji karibu kumbukumbu yote: ongeza kikomo cha kumbukumbu ya GPU na weka context fupi.

Modeli	Toleo	Kumbukumbu	M3 Ultra	M5 Ultra
Mistral Large 3 (675B)675B · 41B zinazotumika	2-bit ubora wa chini	211 GB	29–51	42–75

Kubwa mno 9

Inahitaji kumbukumbu zaidi ya GB 256, hata kwa biti 2.

Maswali

LLM ipi ni bora kwa Mac yenye GB 256?: Miongoni mwa modeli zenye Intelligence Index iliyochapishwa, Qwen3.8 Flash Next ina alama ya juu zaidi kati ya zile zinazoenda vizuri ndani ya GB 256 (6-bit). Kubwa zaidi inayoenda vizuri ni MiMo V2.6 Flash. Kwa code: Qwen3 Coder Next.
Je, GB 256 zinatosha kwa AI ya ndani kwenye Mac?: Modeli huria 54 kati ya 73 zilizo hapa zinaenda vizuri ndani yake kwa biti 4 au zaidi. Kwa chaguo-msingi macOS huiruhusu GPU kutumia takriban GB 192 kati yake; modeli na context yake lazima zitoshe humo.
Je, Mac ya GB 256 inaweza kuendesha gpt-oss-20b?: Ndiyo, 4.25-bit asili, kwa takriban tokeni 194–345 kwa sekunde kwenye M3 Ultra na 284–505 kwenye M5 Ultra.

Ukubwa mwingine: 8 GB · 16 GB · 18 GB · 24 GB · 32 GB · 36 GB · 48 GB · 64 GB · 96 GB · 128 GB · 192 GB · 256 GB · 512 GB

Ziendeshe kwa MLX Harness (bure, open source) au mlx-lm ya Apple ndani ya Terminal: hatua kwa hatua.

Jinsi inavyokokotolewa

Kumbukumbu: vigezo vya modeli kwa biti 8, 6, 4, 3 au 2 kwa kila weight pamoja na group scales za MLX (nusu biti kwa kila weight), au muundo wake asili, pamoja na nafasi ya context juu yake. Kwa chaguo-msingi macOS huiruhusu GPU kutumia takriban 75% ya unified memory. Kasi: bandwidth ya kumbukumbu ikigawanywa kwa baiti zinazosomwa kwa kila tokeni ndiyo kikomo cha juu kwa mazungumzo moja; katika benchmarks za umma za llama.cpp kwenye Apple Silicon uendeshaji halisi hufikia takriban 45% hadi 80% yake, na huo ndio upana unaoonyeshwa. Haya ni makadirio, si vipimo: pima Mac yako mwenyewe kwenye MLX Bench. Takwimu za modeli ni kutoka kwa kadi rasmi ya kila modeli, takwimu za chipu ni kutoka Tech Specs za Apple. Imekaguliwa 2026-10-19. Data ileile ya mlx.app/compare.