Skip to content
Local AI Wizard
Your machine
Theme

Find your next local AI model

Compare open models by quality, memory and license, then see what runs on your computer.

Quality · Overall

0–100 score · higher is better

  • Color = publisher
  • Memory under review
DeepSeek V4.1 Flash · max, DeepSeek81.1 — memory under review
Smaug Agentic, Abacus AI79.5
Kimi K3, Moonshot AI79.2
Qwen3.8 Max, Alibaba78.5
Smaug Flash, Abacus AI77.4 — memory under review
Smaug Mini, Abacus AI76.9
DeepSeek V4 Flash Vision · exp, DeepSeek76.8 — memory under review
Qwen3.8 Flash Next, Alibaba76.2 — memory under review
GLM-5.3, Z.ai76.1
Qwen3.8 27B, Alibaba75.3
GLM-5.2, Z.ai73.2
Inkling · xhigh, Thinking Machines71.9
GLM-5.3 Flash, Z.ai71.6
15 of 23

Source: LiveBench · release of Jun 25, 2026 · checked Oct 18, 2026 · 37 open models without a score in this release are not drawn (no score ≠ 0)

Choose your machine and every chart on this page adapts to its memory.

Intelligence × memory

LiveBench Overall versus estimated memory (log scale) · up and left is better

  • Open model · named when no other is both lighter and better
40608081632641282565121k2k4kMemory needed · model only · GB (log)LiveBench OverallSmaug AgenticQwen3.8 MaxDeepSeek V4 Pro · 0813Smaug Mini

Source: LiveBench · release of Jun 25, 2026 · checked Oct 18, 2026 · memory: Wizard Q4 estimate · not drawn, memory under review: DeepSeek V4.1 Flash · max, Smaug Flash, DeepSeek V4 Flash Vision · exp, and Qwen3.8 Flash Next

Memory needed

Memory each open model needs on its own (model only, before a machine's reserve) · GB, log scale · lower is better

  • Memory under review
1
Llama 3.2 1B InstructLlama 3.2 1B Instruct
2
Qwen3 1.7BQwen3 1.7B
3
Granite 4.2 3BGranite 4.2 3B
3
Llama 3.2 3B InstructLlama 3.2 3B Instruct
3
Ministral 3 3BMinistral 3 3B
3
Phi-4 miniPhi-4 mini
3
Qwen3 4B Instruct 2507Qwen3 4B Instruct 2507
3
SmolLM3 3BSmolLM3 3B
4
Gemma 4 E2BGemma 4 E2B
6
Gemma 4 E4BGemma 4 E4B
6
Granite 4.2 8BGranite 4.2 8B
6
Llama 3.1 8B InstructLlama 3.1 8B Instruct
6
Olmo 3 7B InstructOlmo 3 7B Instruct
7
DeepSeek R1 0528 Qwen3 8BDeepSeek R1 0528 Qw…8B
7
Ministral 3 8BMinistral 3 8B
7
Qwen3 8BQwen3 8B
7
Qwen3.5 9BQwen3.5 9B
9
Gemma 4 12BGemma 4 12B
11
Ministral 3 14BMinistral 3 14B
11
Phi-4Phi-4
12
Qwen3 14BQwen3 14B
16
gpt-oss-20bgpt-oss-20b
18
Devstral Small 2 24BDevstral Small 2 24B
19
Gemma 4 26B A4BGemma 4 26B A4B
19
Mistral Small 3.2 24BMistral Small 3.2 24B
21
Qwen3.6 27BQwen3.6 27B
21
Qwen3.8 27BQwen3.8 27B
21
Smaug MiniSmaug Mini
23
GLM-4.7 FlashGLM-4.7 Flash
23
Granite 4.2 30BGranite 4.2 30B
23
Nemotron 3.5 Lightning 30B A3BNemotron 3.5 Lightnin…A3B
23
Qwen3 30B A3B Instruct 2507Qwen3 30B A3B In…2507
23
Qwen3 Coder 30B A3BQwen3 Coder 30B A3B
24
Gemma 4 31BGemma 4 31B
24
Olmo 3 32B ThinkOlmo 3 32B Think
27
Qwen3.5 35B A3BQwen3.5 35B A3B
30
Nemotron 3 Nano 30B A3BNemotron 3 Nano 30B A3B
90
Nemotron 3 Super 120B A12BNemotron 3 Super 1…A12B
93
Ling 3.0 FlashLing 3.0 Flash
213
DeepSeek V4 FlashDeepSeek V4 Flash
228
DeepSeek V4 Flash · 0731DeepSeek V4 Flash · 0731
232
MiMo V2.6 FlashMiMo V2.6 Flash
240
GLM-5.3 FlashGLM-5.3 Flash
321
MiniMax M3MiniMax M3
413
Nemotron 3 Ultra 550B A55BNemotron 3 Ultra 55…A55B
558
GLM-5.2GLM-5.2
558
GLM-5.3GLM-5.3
732
Inkling · xhighInkling · xhigh
750
Kimi K2.6 · thinkingKimi K2.6 · thinking
750
Kimi K2.7 CodeKimi K2.7 Code
765
MiMo V2.6 ProMiMo V2.6 Pro
1,200
DeepSeek V4 ProDeepSeek V4 Pro
1,200
DeepSeek V4 Pro · 0813DeepSeek V4 Pro · 0813
1,800
Qwen3.8 MaxQwen3.8 Max
2,100
Kimi K3Kimi K3
2,100
Smaug AgenticSmaug Agentic
—
DeepSeek V4 Flash Vision · expDeepSeek V4 Flash Vi…exp
—
DeepSeek V4.1 Flash · maxDeepSeek V4.1 Flash · max
—
Qwen3.8 Flash NextQwen3.8 Flash Next
—
Smaug FlashSmaug Flash

Source: Wizard estimate from published parameter counts

Licenses

45 of 60 open models use a permissive license (MIT or Apache 2.0); the others use the publisher's own terms

Apache 2.0, Qwen3.5 9B, Qwen3.8 27B, Qwen3.5 35B A3B, Qwen3.6 27B, Inkling · xhigh, Smaug Mini, Qwen3 1.7B, Qwen3 4B Instruct 2507, Qwen3 8B, Qwen3 14B, Qwen3 30B A3B Instruct 2507, Qwen3 Coder 30B A3B, Gemma 4 E2B, Gemma 4 E4B, Gemma 4 12B, Gemma 4 26B A4B, Gemma 4 31B, gpt-oss-20b, Mistral Small 3.2 24B, Ministral 3 3B, Ministral 3 8B, Ministral 3 14B, Devstral Small 2 24B, Granite 4.2 3B, Granite 4.2 8B, Granite 4.2 30B, Olmo 3 7B Instruct, Olmo 3 32B Think, SmolLM3 3B29
MIT, DeepSeek V4 Flash · 0731, GLM-5.3 Flash, DeepSeek V4.1 Flash · max, DeepSeek V4 Flash, DeepSeek V4 Pro, GLM-5.2, DeepSeek V4 Pro · 0813, DeepSeek V4 Flash Vision · exp, Smaug Flash, MiMo V2.6 Pro, MiMo V2.6 Flash, Ling 3.0 Flash, Phi-4 mini, Phi-4, GLM-4.7 Flash, DeepSeek R1 0528 Qwen3 8B16
Other publisher terms, GLM-5.3, MiniMax M3, Qwen3.8 Max, Qwen3.8 Flash Next, Llama 3.1 8B Instruct5
OpenMDW 1.1, Nemotron 3 Ultra 550B A55B, Nemotron 3.5 Lightning 30B A3B2
Publisher terms: Kimi K3, Kimi K3, Smaug Agentic2
Publisher terms: Llama 3.2, Llama 3.2 3B Instruct, Llama 3.2 1B Instruct2
Publisher terms: Modified MIT, Kimi K2.6 · thinking, Kimi K2.7 Code2
Publisher terms: NVIDIA Nemotron Open Model License, Nemotron 3 Super 120B A12B, Nemotron 3 Nano 30B A3B2

Source: each model's public card on Hugging Face

By task

Only text models have a comparable score so far; the other tasks show what is catalogued.

Compare models

Quality, memory and license side by side. Up to 3 models.

20 published comparisonsOpen comparison