LLM ipi ni bora kwa M5 Ultra yenye GB 512?

Miongoni mwa modeli zenye Intelligence Index iliyochapishwa, DeepSeek V4.1 Flash 1017 ina alama ya juu zaidi kati ya zile zinazoenda vizuri hapa: 4-bit, takriban tokeni 60–107 kwa sekunde.

Je, M5 Ultra yenye GB 512 inaweza kuendesha Llama 3.3 70B?

Ndiyo: 4-bit, kwa takriban tokeni 14–24 kwa sekunde.

gpt-oss-20b ina kasi gani kwenye M5 Ultra?

Takriban tokeni 284–505 kwa sekunde (4.25-bit asili), kutokana na bandwidth ya kumbukumbu ya chipu ya 1,200 GB/s.

M5 Ultra yenye GB 512 inaendesha LLM zipi? Modeli na kasi

Zinaenda vizuri 68

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

Modeli	Toleo	Kumbukumbu	Tokeni/s
DeepSeek V4.1 Flash 1017552B · 16B zinazotumika · Akili 45.1	4-bit	311 GB	60–107
GLM-5.3 Flash320B · 18B zinazotumika · Akili 41.8	8-bit	340 GB	28–50
Qwen3.8 Flash Next180B · 6B zinazotumika · Akili 39.8	8-bit	191 GB	84–150
MiMo V2.6 Flash309B · 15B zinazotumika · Akili 37.9	4.48-bit asili	173 GB	64–114
Qwen3.8 27B27B · Akili 33.7	8-bit	28.7 GB	19–33
Beam 1.0501B · 23B zinazotumika · Akili 30.3	4-bit	282 GB	42–74
MiniMax M3428B · 23B zinazotumika · Akili 29.2	6-bit	348 GB	29–51
Inkling Small276B · 12B zinazotumika · Akili 25.7	8-bit	293 GB	42–75
Hy3295B · 21B zinazotumika · Akili 25.3	8-bit	313 GB	24–43
Nemotron 3 Ultra 550B A55B550B · 55B zinazotumika · Akili 22.9	4-bit	309 GB	17–31
Nemotron 3.5 Bolt 74B A9B74B · 9B zinazotumika · Akili 21.4	8-bit	78.6 GB	56–100
Qwen3.5 397B A17B397B · 17B zinazotumika · Akili 18.4	6-bit	323 GB	39–70
Qwen3.6 35B A3B35B · 3B zinazotumika · Akili 18.2	8-bit	37.2 GB	169–300
Muse Glimmer 30B29.6B · Akili 17.5	8-bit	31.5 GB	17–30
Qwen3.5 122B A10B122B · 10B zinazotumika · Akili 15.6	8-bit	130 GB	51–91
Gemma 4 31B30.7B · Akili 14.7	8-bit	32.6 GB	17–29
Nemotron 3.5 Lightning 30B A3B30B · 3B zinazotumika · Akili 12.9	8-bit	31.9 GB	169–300
Granite 4.2 30B30B · Akili 12.7	8-bit	31.9 GB	17–30
MiniCPM5 2B2.52B · Akili 12.5	8-bit	2.7 GB	200–356
gpt-oss-120b117B · 5.1B zinazotumika · Akili 11.6	4.25-bit asili	65.3 GB	200–356
Mistral Small 4 (119B)119B · 6.5B zinazotumika · Akili 11.3	8-bit	126 GB	78–139
Qwen3.5 9B9B · Akili 11.2	8-bit	9.6 GB	56–100
Granite 4.2 8B8B · Akili 11.1	8-bit	8.5 GB	64–113
Llama 4 Maverick 17B-128E Instruct400B · 17B zinazotumika · Akili 9.3	6-bit	325 GB	39–70
Qwen3 Coder Next80B · 3B zinazotumika · Akili 9.2	8-bit	85 GB	169–300
Granite 4.2 3B3B · Akili 9.1	8-bit	3.2 GB	169–300
gpt-oss-20b21B · 3.6B zinazotumika · Akili 9	4.25-bit asili	13.8 GB	284–505
Nemotron 3 Nano 30B A3B30B · 3.5B zinazotumika · Akili 8.9	8-bit	31.9 GB	146–259
Gemma 4 E4B8B · Akili 8.9	8-bit	8.5 GB	113–200
Llama 4 Scout 17B-16E Instruct109B · 17B zinazotumika · Akili 6.5	8-bit	116 GB	30–53
Ministral 3 8B8.9B · Akili 5.5	8-bit	9.5 GB	57–101
Nemotron 3 Nano 4B3.97B · Akili 4.8	8-bit	4.2 GB	129–229
DeepSeek V4 Flash 0731284B · 13B zinazotumika	8-bit	302 GB	39–70
MiniMax M2.7230B · 10B zinazotumika	8-bit	244 GB	51–91
Ornith 1.5 122B122B · 10B zinazotumika	8-bit	130 GB	51–91
Nemotron 3 Super 120B A12B120B · 12B zinazotumika	8-bit	128 GB	42–75
GLM-4.5 Air106B · 12B zinazotumika	8-bit	113 GB	42–75
Llama 3.3 70B Instruct70B	4-bit	39.4 GB	14–24
Qwen3 Coder 30B A3B Instruct30.5B · 3.3B zinazotumika	8-bit	32.4 GB	154–274
GLM-4.7 Flash30B · 3B zinazotumika	8-bit	31.9 GB	169–300
Qwen3.6 27B27B	8-bit	28.7 GB	19–33
Gemma 4 26B A4B25.2B · 3.8B zinazotumika	8-bit	26.8 GB	135–240
Devstral Small 2 (24B)24B	8-bit	25.5 GB	21–38
Magistral Small 1.2 (2509)24B	8-bit	25.5 GB	21–38
Mistral Small 3.2 (24B)24B	8-bit	25.5 GB	21–38
LFM2.5 24B A2B24B · 2B zinazotumika	8-bit	25.5 GB	257–457
Phi-4 Reasoning Vision 15B15B	8-bit	15.9 GB	34–60
Phi-4 (14B)14B	8-bit	14.9 GB	36–64
Gemma 4 12B Unified12B	8-bit	12.7 GB	43–76
Qwen3 8B8.2B	8-bit	8.7 GB	62–110
Llama 3.1 8B Instruct8B	8-bit	8.5 GB	64–113
Gemma 4 E2B5.1B	8-bit	5.4 GB	225–400
Qwen3.5 4B4B	8-bit	4.3 GB	126–223
Phi-4 mini instruct3.8B	8-bit	4 GB	135–240
Llama 3.2 3B Instruct3.21B	8-bit	3.4 GB	159–282
SmolLM3 3B3B	8-bit	3.2 GB	169–300
LFM2.5 2.6B2.69B	8-bit	2.9 GB	186–331
Qwen3.5 2B2B	8-bit	2.1 GB	257–457
Llama 3.2 1B Instruct1.23B	8-bit	1.3 GB	415–738
Qwen3.5 0.8B0.8B	8-bit	0.9 GB	600–1067
Qwen3 0.6B0.6B	8-bit	0.6 GB	900–1600
MiMo V2.6 Pro1,020B · 42B zinazotumika · Akili 46.3	2-bit ubora wa chini	319 GB	41–73
GLM-5.3744B · 40B zinazotumika · Akili 44.8	3-bit ubora wa chini	326 GB	31–55
Muse Spark Open816B · 42B zinazotumika · Akili 38.9	2-bit ubora wa chini	255 GB	41–73
Kimi K2.7 Code1,000B · 32B zinazotumika · Akili 25.8	2-bit ubora wa chini	313 GB	54–96
Inkling975B · 41B zinazotumika · Akili 25	2-bit ubora wa chini	305 GB	42–75
Kimi K2.61,000B · 32B zinazotumika	2-bit ubora wa chini	313 GB	54–96
Mistral Large 3 (675B)675B · 41B zinazotumika	3-bit ubora wa chini	295 GB	30–54

Zinaenda, lakini polepole 2

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

Modeli	Toleo	Kumbukumbu	Tokeni/s
Mistral Medium 3.5 (128B)128B · Akili 14.2	8-bit	136 GB	4–7
Devstral 2 (123B)123B	8-bit	131 GB	4–7

Kubwa mno 3

Inahitaji kumbukumbu zaidi ya GB 512, hata kwa biti 2.

Mac nyingine

Kila Mac yenye GB 512, ya chipu yoyote.

M5 Ultra yenye kumbukumbu zaidi au kidogo:

Chipu jirani:

M5 Max 128 GB47 zinaenda vizuri

Mac Studio M5 Ultra

Ana kwa ana:

Dhidi ya PC na kadi za michoro:

Ilinganishe na Mac au GPU yoyote · M5 Ultra kwa AI ya ndani ya kifaa · Mac bora kwa LLM za ndani · Kila Mac

Maswali

LLM ipi ni bora kwa M5 Ultra yenye GB 512?: Miongoni mwa modeli zenye Intelligence Index iliyochapishwa, DeepSeek V4.1 Flash 1017 ina alama ya juu zaidi kati ya zile zinazoenda vizuri hapa: 4-bit, takriban tokeni 60–107 kwa sekunde.
Je, M5 Ultra yenye GB 512 inaweza kuendesha Llama 3.3 70B?: Ndiyo: 4-bit, kwa takriban tokeni 14–24 kwa sekunde.
gpt-oss-20b ina kasi gani kwenye M5 Ultra?: Takriban tokeni 284–505 kwa sekunde (4.25-bit asili), kutokana na bandwidth ya kumbukumbu ya chipu ya 1,200 GB/s.

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.