Hardware HardwareSystemActive

MacBook Air (Apple M5)

Apple spec sheet data quality79

Memory: 16–32 GB
Bandwidth: 153 GB/s
TDP: —

Run locally with this device →

Specifications

As published

Every value opens its evidence (source, tier, observed time).

Kind: Source:Apple — Mac tech specsT1observed 11 h agohigh
Kind raw: computer
Spec sheet: apple.com/macbook-air/specs
Source:Apple — Mac tech specsT1observed 11 h agohigh
Memory: Source:Apple — Mac tech specsT1observed 11 h agohigh
Memory type: Source:Apple — Mac tech specsT1observed 11 h agohigh
Product line: Source:Apple — Mac tech specsT1observed 11 h agohigh
Memory bandwidth: Source:Apple — Mac tech specsT1observed 11 h agohigh

What can this run?

Models estimated to fit Estimated

Estimate from each model's parameter count, the chosen quantization and context — never a measurement.

Open in Run locally →

Quantization4bit 8bit fp16

Context8.19K 32.8K 131.1K

Configuration16 GB 24 GB 32 GB

EstimatedAll fit figures are estimates, not measurements.341 of 536 evaluated models fit

Assumptions (7)

Estimated, not measured: weights = parameters × bytes/param × 1.15 runtime overhead (or the observed artifact file size when one is recorded).
bytes/param: 4bit = 0.5, 8bit = 1.0, fp16 = 2.0 (uniform quantization, no per-layer exceptions).
KV cache: 2 × layers × kv_heads × head_dim × 2 bytes × context × batch when the architecture is known; otherwise 0.5 GB per 8 192 tokens (× batch), independent of architecture (GQA/MLA models need less).
A model 'fits' when the estimate is at most the device memory minus 2 GB reserved for the OS and framework.
Mixture-of-experts models are estimated on total parameters (all experts must be resident); active parameters are ignored.
Device memory uses the largest configuration when several are listed (e.g. Apple silicon tiers).
Multi-GPU: device memories are summed; interconnect bandwidth, tensor-parallel replication and pipeline bubbles are not modelled.

This device ships in 3 memory configurations (16 / 24 / 32 GB). Estimates use 24 GB — pick another configuration above. Multiple units: use Run locally (memories are summed; interconnect is not modelled).

Model	Params	Est. memory	Headroom	Fits	Breakdown
VibeVoice-ASROpen weightsMicrosoft	8.67B	21.9 GB estimated	+0.1 GB	✓ fits	weights 17.4 GB estimated KV cache 2.00 GB (heuristic) overhead 2.6 GB reserved 2 GB context 32,768 × batch 1
Molmo2-8BOpen weightsAllen Institute for AI	8.66B	21.9 GB estimated	+0.1 GB	✓ fits	weights 17.3 GB estimated KV cache 2.00 GB (heuristic) overhead 2.6 GB reserved 2 GB context 32,768 × batch 1
aya-vision-8brestricted-weightsCohere	8.63B	21.9 GB estimated	+0.1 GB	✓ fits	weights 17.3 GB estimated KV cache 2.00 GB (heuristic) overhead 2.6 GB reserved 2 GB context 32,768 × batch 1
Intern-S1-miniOpen weightsInternLM (Shanghai AI Laboratory)	8.54B	21.6 GB estimated	+0.4 GB	✓ fits	weights 17.1 GB estimated KV cache 2.00 GB (heuristic) overhead 2.6 GB reserved 2 GB context 32,768 × batch 1
GKA-primed-HQwen3-8B-ReasonerOpen weightsAmazon Web Services	8.5B	21.6 GB estimated	+0.5 GB	✓ fits	weights 17.0 GB estimated KV cache 2.00 GB (heuristic) overhead 2.5 GB reserved 2 GB context 32,768 × batch 1
GDN-primed-HQwen3-8B-InstructOpen weightsAmazon Web Services	8.5B	21.6 GB estimated	+0.5 GB	✓ fits	weights 17.0 GB estimated KV cache 2.00 GB (heuristic) overhead 2.5 GB reserved 2 GB context 32,768 × batch 1
LFM2.5-8B-A1BOpen weightsLiquid AI	8.47B	21.5 GB estimated	+0.5 GB	✓ fits	weights 16.9 GB estimated KV cache 2.00 GB (heuristic) overhead 2.5 GB reserved 2 GB context 32,768 × batch 1
Qwen2.5-VL-7B-InstructOpen weightsQwen	8.29B	21.1 GB estimated	+0.9 GB	✓ fits	weights 16.6 GB estimated KV cache 2.00 GB (heuristic) overhead 2.5 GB reserved 2 GB context 32,768 × batch 1
UI-TARS 7BOpen weightsByteDance	8.29B	21.1 GB estimated	+0.9 GB	✓ fits	weights 16.6 GB estimated KV cache 2.00 GB (heuristic) overhead 2.5 GB reserved 2 GB context 32,768 × batch 1
olmOCR-2-7B-1025Open weightsAllen Institute for AI	8.29B	21.1 GB estimated	+0.9 GB	✓ fits	weights 16.6 GB estimated KV cache 2.00 GB (heuristic) overhead 2.5 GB reserved 2 GB context 32,768 × batch 1
UI-TARS-7B-SFTOpen weightsByteDance	8.29B	21.1 GB estimated	+0.9 GB	✓ fits	weights 16.6 GB estimated KV cache 2.00 GB (heuristic) overhead 2.5 GB reserved 2 GB context 32,768 × batch 1
UI-TARS-7B-DPOOpen weightsByteDance	8.29B	21.1 GB estimated	+0.9 GB	✓ fits	weights 16.6 GB estimated KV cache 2.00 GB (heuristic) overhead 2.5 GB reserved 2 GB context 32,768 × batch 1
Seed-Coder-8B-ReasoningOpen weightsByteDance	8.25B	21.0 GB estimated	+1.0 GB	✓ fits	weights 16.5 GB estimated KV cache 2.00 GB (heuristic) overhead 2.5 GB reserved 2 GB context 32,768 × batch 1
Seed-Coder-8B-InstructOpen weightsByteDance	8.25B	21.0 GB estimated	+1.0 GB	✓ fits	weights 16.5 GB estimated KV cache 2.00 GB (heuristic) overhead 2.5 GB reserved 2 GB context 32,768 × batch 1
Stable-DiffCoder-8B-InstructOpen weightsByteDance	8.25B	21.0 GB estimated	+1.0 GB	✓ fits	weights 16.5 GB estimated KV cache 2.00 GB (heuristic) overhead 2.5 GB reserved 2 GB context 32,768 × batch 1
Stable-DiffCoder-8B-BaseOpen weightsByteDance	8.25B	21.0 GB estimated	+1.0 GB	✓ fits	weights 16.5 GB estimated KV cache 2.00 GB (heuristic) overhead 2.5 GB reserved 2 GB context 32,768 × batch 1
Seed-Coder-8B-BaseOpen weightsByteDance	8.25B	21.0 GB estimated	+1.0 GB	✓ fits	weights 16.5 GB estimated KV cache 2.00 GB (heuristic) overhead 2.5 GB reserved 2 GB context 32,768 × batch 1
DeepSeek-R1-0528-Qwen3-8BOpen weightsDeepSeek	8.19B	20.8 GB estimated	+1.2 GB	✓ fits	weights 16.4 GB estimated KV cache 2.00 GB (heuristic) overhead 2.5 GB reserved 2 GB context 32,768 × batch 1
Qwen3 8BOpen weightsQwen	8.19B	20.8 GB estimated	+1.2 GB	✓ fits	weights 16.4 GB estimated KV cache 2.00 GB (heuristic) overhead 2.5 GB reserved 2 GB context 32,768 × batch 1
granite-guardian-3.3-8bOpen weightsIBM	8.17B	20.8 GB estimated	+1.2 GB	✓ fits	weights 16.3 GB estimated KV cache 2.00 GB (heuristic) overhead 2.5 GB reserved 2 GB context 32,768 × batch 1
granite-3.0-8b-instructOpen weightsIBM	8.17B	20.8 GB estimated	+1.2 GB	✓ fits	weights 16.3 GB estimated KV cache 2.00 GB (heuristic) overhead 2.5 GB reserved 2 GB context 32,768 × batch 1
stable-diffusion-3.5-largeOpen weightsStability AI	8.15B	20.7 GB estimated	+1.3 GB	✓ fits	weights 16.3 GB estimated KV cache 2.00 GB (heuristic) overhead 2.4 GB reserved 2 GB context 32,768 × batch 1
ERNIE-Image-TurboOpen weightsBaidu	8.03B	20.5 GB estimated	+1.5 GB	✓ fits	weights 16.1 GB estimated KV cache 2.00 GB (heuristic) overhead 2.4 GB reserved 2 GB context 32,768 × batch 1
ERNIE-ImageOpen weightsBaidu	8.03B	20.5 GB estimated	+1.5 GB	✓ fits	weights 16.1 GB estimated KV cache 2.00 GB (heuristic) overhead 2.4 GB reserved 2 GB context 32,768 × batch 1
Meta-Llama-3-8Brestricted-weightsMeta AI	8.03B	20.5 GB estimated	+1.5 GB	✓ fits	weights 16.1 GB estimated KV cache 2.00 GB (heuristic) overhead 2.4 GB reserved 2 GB context 32,768 × batch 1
NousResearch/Meta-Llama-3-8B-InstructOpen weightsNous Research	8.03B	20.5 GB estimated	+1.5 GB	✓ fits	weights 16.1 GB estimated KV cache 2.00 GB (heuristic) overhead 2.4 GB reserved 2 GB context 32,768 × batch 1
Hermes-2-Theta-Llama-3-8BOpen weightsNous Research	8.03B	20.5 GB estimated	+1.5 GB	✓ fits	weights 16.1 GB estimated KV cache 2.00 GB (heuristic) overhead 2.4 GB reserved 2 GB context 32,768 × batch 1
NousResearch/Meta-Llama-3.1-8Brestricted-weightsNous Research	8.03B	20.5 GB estimated	+1.5 GB	✓ fits	weights 16.1 GB estimated KV cache 2.00 GB (heuristic) overhead 2.4 GB reserved 2 GB context 32,768 × batch 1
Llama 3.1 8Brestricted-weightsMeta AI	8.03B	20.5 GB estimated	+1.5 GB	✓ fits	weights 16.1 GB estimated KV cache 2.00 GB (heuristic) overhead 2.4 GB reserved 2 GB context 32,768 × batch 1
NousResearch/Meta-Llama-3-8BOpen weightsNous Research	8.03B	20.5 GB estimated	+1.5 GB	✓ fits	weights 16.1 GB estimated KV cache 2.00 GB (heuristic) overhead 2.4 GB reserved 2 GB context 32,768 × batch 1
Hermes-3-Llama-3.1-8Brestricted-weightsNous Research	8.03B	20.5 GB estimated	+1.5 GB	✓ fits	weights 16.1 GB estimated KV cache 2.00 GB (heuristic) overhead 2.4 GB reserved 2 GB context 32,768 × batch 1
NousResearch/Meta-Llama-3.1-8B-Instructrestricted-weightsNous Research	8.03B	20.5 GB estimated	+1.5 GB	✓ fits	weights 16.1 GB estimated KV cache 2.00 GB (heuristic) overhead 2.4 GB reserved 2 GB context 32,768 × batch 1
Llama 3.1 8B Instructrestricted-weightsMeta AI	8.03B	20.5 GB estimated	+1.5 GB	✓ fits	weights 16.1 GB estimated KV cache 2.00 GB (heuristic) overhead 2.4 GB reserved 2 GB context 32,768 × batch 1
Salesforce/Llama-xLAM-2-8b-fc-rrestricted-weightsSalesforce	8.03B	20.5 GB estimated	+1.5 GB	✓ fits	weights 16.1 GB estimated KV cache 2.00 GB (heuristic) overhead 2.4 GB reserved 2 GB context 32,768 × batch 1
Meta-Llama-3-8B-Instructrestricted-weightsMeta AI	8.03B	20.5 GB estimated	+1.5 GB	✓ fits	weights 16.1 GB estimated KV cache 2.00 GB (heuristic) overhead 2.4 GB reserved 2 GB context 32,768 × batch 1
Hy-MT2-7BOpen weightsTencent	8.03B	20.5 GB estimated	+1.5 GB	✓ fits	weights 16.1 GB estimated KV cache 2.00 GB (heuristic) overhead 2.4 GB reserved 2 GB context 32,768 × batch 1
aya-23-8Brestricted-weightsCohere	8.03B	20.5 GB estimated	+1.5 GB	✓ fits	weights 16.1 GB estimated KV cache 2.00 GB (heuristic) overhead 2.4 GB reserved 2 GB context 32,768 × batch 1
aya-expanse-8brestricted-weightsCohere	8.03B	20.5 GB estimated	+1.5 GB	✓ fits	weights 16.1 GB estimated KV cache 2.00 GB (heuristic) overhead 2.4 GB reserved 2 GB context 32,768 × batch 1
c4ai-command-r7b-12-2024restricted-weightsCohere	8.03B	20.5 GB estimated	+1.5 GB	✓ fits	weights 16.1 GB estimated KV cache 2.00 GB (heuristic) overhead 2.4 GB reserved 2 GB context 32,768 × batch 1
Ministral-8B-Instruct-2410Open weightsMistral AI	8.02B	20.4 GB estimated	+1.6 GB	✓ fits	weights 16.0 GB estimated KV cache 2.00 GB (heuristic) overhead 2.4 GB reserved 2 GB context 32,768 × batch 1
Robostral NavigateProprietaryMistral AI	8B	20.4 GB estimated	+1.6 GB	✓ fits	weights 16.0 GB estimated KV cache 2.00 GB (heuristic) overhead 2.4 GB reserved 2 GB context 32,768 × batch 1
gemma-4-E4B-itOpen weightsGoogle	8B	20.4 GB estimated	+1.6 GB	✓ fits	weights 16.0 GB estimated KV cache 2.00 GB (heuristic) overhead 2.4 GB reserved 2 GB context 32,768 × batch 1
ERNIE-Image-AesOpen weightsBaidu	7.94B	20.3 GB estimated	+1.7 GB	✓ fits	weights 15.9 GB estimated KV cache 2.00 GB (heuristic) overhead 2.4 GB reserved 2 GB context 32,768 × batch 1
internlm2_5-7b-chatOpen weightsInternLM (Shanghai AI Laboratory)	7.74B	19.8 GB estimated	+2.2 GB	✓ fits	weights 15.5 GB estimated KV cache 2.00 GB (heuristic) overhead 2.3 GB reserved 2 GB context 32,768 × batch 1
internlm2-chat-7bOpen weightsInternLM (Shanghai AI Laboratory)	7.74B	19.8 GB estimated	+2.2 GB	✓ fits	weights 15.5 GB estimated KV cache 2.00 GB (heuristic) overhead 2.3 GB reserved 2 GB context 32,768 × batch 1
StripedHyena-Hessian-7BOpen weightsTogether AI	7.65B	19.6 GB estimated	+2.4 GB	✓ fits	weights 15.3 GB estimated KV cache 2.00 GB (heuristic) overhead 2.3 GB reserved 2 GB context 32,768 × batch 1
StripedHyena-Nous-7BOpen weightsTogether AI	7.65B	19.6 GB estimated	+2.4 GB	✓ fits	weights 15.3 GB estimated KV cache 2.00 GB (heuristic) overhead 2.3 GB reserved 2 GB context 32,768 × batch 1
SynLogic-7BOpen weightsMiniMax	7.62B	19.5 GB estimated	+2.5 GB	✓ fits	weights 15.2 GB estimated KV cache 2.00 GB (heuristic) overhead 2.3 GB reserved 2 GB context 32,768 × batch 1
Qwen2.5 7B InstructOpen weightsQwen	7.62B	19.5 GB estimated	+2.5 GB	✓ fits	weights 15.2 GB estimated KV cache 2.00 GB (heuristic) overhead 2.3 GB reserved 2 GB context 32,768 × batch 1
BFS-Prover-V2-7BOpen weightsByteDance	7.62B	19.5 GB estimated	+2.5 GB	✓ fits	weights 15.2 GB estimated KV cache 2.00 GB (heuristic) overhead 2.3 GB reserved 2 GB context 32,768 × batch 1
DeepSeek-R1-Distill-Qwen-7BOpen weightsDeepSeek	7.62B	19.5 GB estimated	+2.5 GB	✓ fits	weights 15.2 GB estimated KV cache 2.00 GB (heuristic) overhead 2.3 GB reserved 2 GB context 32,768 × batch 1
Seed-X-PPO-7BOpen weightsByteDance	7.51B	19.3 GB estimated	+2.7 GB	✓ fits	weights 15.0 GB estimated KV cache 2.00 GB (heuristic) overhead 2.3 GB reserved 2 GB context 32,768 × batch 1
Hunyuan-7B-InstructOpen weightsTencent	7.5B	19.3 GB estimated	+2.7 GB	✓ fits	weights 15.0 GB estimated KV cache 2.00 GB (heuristic) overhead 2.3 GB reserved 2 GB context 32,768 × batch 1
OLMo-2-1124-7B-InstructOpen weightsAllen Institute for AI	7.3B	18.8 GB estimated	+3.2 GB	✓ fits	weights 14.6 GB estimated KV cache 2.00 GB (heuristic) overhead 2.2 GB reserved 2 GB context 32,768 × batch 1
OLMo-2-1124-7BOpen weightsAllen Institute for AI	7.3B	18.8 GB estimated	+3.2 GB	✓ fits	weights 14.6 GB estimated KV cache 2.00 GB (heuristic) overhead 2.2 GB reserved 2 GB context 32,768 × batch 1
Olmo-3-7B-InstructOpen weightsAllen Institute for AI	7.3B	18.8 GB estimated	+3.2 GB	✓ fits	weights 14.6 GB estimated KV cache 2.00 GB (heuristic) overhead 2.2 GB reserved 2 GB context 32,768 × batch 1
Olmo-3-7B-ThinkOpen weightsAllen Institute for AI	7.3B	18.8 GB estimated	+3.2 GB	✓ fits	weights 14.6 GB estimated KV cache 2.00 GB (heuristic) overhead 2.2 GB reserved 2 GB context 32,768 × batch 1
Olmo-3-1025-7BOpen weightsAllen Institute for AI	7.3B	18.8 GB estimated	+3.2 GB	✓ fits	weights 14.6 GB estimated KV cache 2.00 GB (heuristic) overhead 2.2 GB reserved 2 GB context 32,768 × batch 1
wildguardOpen weightsAllen Institute for AI	7.25B	18.7 GB estimated	+3.3 GB	✓ fits	weights 14.5 GB estimated KV cache 2.00 GB (heuristic) overhead 2.2 GB reserved 2 GB context 32,768 × batch 1
Mistral-7B-v0.3Open weightsMistral AI	7.25B	18.7 GB estimated	+3.3 GB	✓ fits	weights 14.5 GB estimated KV cache 2.00 GB (heuristic) overhead 2.2 GB reserved 2 GB context 32,768 × batch 1
Mistral-7B-Instruct-v0.3Open weightsMistral AI	7.25B	18.7 GB estimated	+3.3 GB	✓ fits	weights 14.5 GB estimated KV cache 2.00 GB (heuristic) overhead 2.2 GB reserved 2 GB context 32,768 × batch 1
Mistral-7B-v0.1Open weightsMistral AI	7.24B	18.6 GB estimated	+3.4 GB	✓ fits	weights 14.5 GB estimated KV cache 2.00 GB (heuristic) overhead 2.2 GB reserved 2 GB context 32,768 × batch 1
Mistral-7B-Instruct-v0.2Open weightsMistral AI	7.24B	18.6 GB estimated	+3.4 GB	✓ fits	weights 14.5 GB estimated KV cache 2.00 GB (heuristic) overhead 2.2 GB reserved 2 GB context 32,768 × batch 1
Mistral-7B-Instruct-v0.1Open weightsMistral AI	7.24B	18.6 GB estimated	+3.4 GB	✓ fits	weights 14.5 GB estimated KV cache 2.00 GB (heuristic) overhead 2.2 GB reserved 2 GB context 32,768 × batch 1
SFR-Embedding-2_Rrestricted-weightsSalesforce	7.11B	18.4 GB estimated	+3.6 GB	✓ fits	weights 14.2 GB estimated KV cache 2.00 GB (heuristic) overhead 2.1 GB reserved 2 GB context 32,768 × batch 1
SFR-Embedding-Mistralrestricted-weightsSalesforce	7.11B	18.4 GB estimated	+3.6 GB	✓ fits	weights 14.2 GB estimated KV cache 2.00 GB (heuristic) overhead 2.1 GB reserved 2 GB context 32,768 × batch 1
SeedVR-7BOpen weightsByteDance	7B	18.1 GB estimated	+3.9 GB	✓ fits	weights 14.0 GB estimated KV cache 2.00 GB (heuristic) overhead 2.1 GB reserved 2 GB context 32,768 × batch 1
internlm-xcomposer-7bOpen weightsInternLM (Shanghai AI Laboratory)	7B	18.1 GB estimated	+3.9 GB	✓ fits	weights 14.0 GB estimated KV cache 2.00 GB (heuristic) overhead 2.1 GB reserved 2 GB context 32,768 × batch 1
internlm2-base-7bOpen weightsInternLM (Shanghai AI Laboratory)	7B	18.1 GB estimated	+3.9 GB	✓ fits	weights 14.0 GB estimated KV cache 2.00 GB (heuristic) overhead 2.1 GB reserved 2 GB context 32,768 × batch 1
internlm-xcomposer2-7bOpen weightsInternLM (Shanghai AI Laboratory)	7B	18.1 GB estimated	+3.9 GB	✓ fits	weights 14.0 GB estimated KV cache 2.00 GB (heuristic) overhead 2.1 GB reserved 2 GB context 32,768 × batch 1
Qwen2-AudioOpen weightsQwen Team	7B	18.1 GB estimated	+3.9 GB	✓ fits	weights 14.0 GB estimated KV cache 2.00 GB (heuristic) overhead 2.1 GB reserved 2 GB context 32,768 × batch 1
SeedVR2-7BOpen weightsByteDance	7B	18.1 GB estimated	+3.9 GB	✓ fits	weights 14.0 GB estimated KV cache 2.00 GB (heuristic) overhead 2.1 GB reserved 2 GB context 32,768 × batch 1
fireworks-ai/mistral-7b-eagle-head-experimentalOpen weightsFireworks AI	7B	18.1 GB estimated	+3.9 GB	✓ fits	weights 14.0 GB estimated KV cache 2.00 GB (heuristic) overhead 2.1 GB reserved 2 GB context 32,768 × batch 1
internlm-chat-7bOpen weightsInternLM (Shanghai AI Laboratory)	7B	18.1 GB estimated	+3.9 GB	✓ fits	weights 14.0 GB estimated KV cache 2.00 GB (heuristic) overhead 2.1 GB reserved 2 GB context 32,768 × batch 1
internlm2-7bOpen weightsInternLM (Shanghai AI Laboratory)	7B	18.1 GB estimated	+3.9 GB	✓ fits	weights 14.0 GB estimated KV cache 2.00 GB (heuristic) overhead 2.1 GB reserved 2 GB context 32,768 × batch 1
togethercomputer/LLaMA-2-7B-32Krestricted-weightsTogether AI	7B	18.1 GB estimated	+3.9 GB	✓ fits	weights 14.0 GB estimated KV cache 2.00 GB (heuristic) overhead 2.1 GB reserved 2 GB context 32,768 × batch 1
RedPajama-INCITE-7B-BaseOpen weightsTogether AI	7B	18.1 GB estimated	+3.9 GB	✓ fits	weights 14.0 GB estimated KV cache 2.00 GB (heuristic) overhead 2.1 GB reserved 2 GB context 32,768 × batch 1
togethercomputer/Llama-2-7B-32K-Instructrestricted-weightsTogether AI	7B	18.1 GB estimated	+3.9 GB	✓ fits	weights 14.0 GB estimated KV cache 2.00 GB (heuristic) overhead 2.1 GB reserved 2 GB context 32,768 × batch 1
RedPajama-INCITE-7B-ChatOpen weightsTogether AI	7B	18.1 GB estimated	+3.9 GB	✓ fits	weights 14.0 GB estimated KV cache 2.00 GB (heuristic) overhead 2.1 GB reserved 2 GB context 32,768 × batch 1
RedPajama-INCITE-7B-InstructOpen weightsTogether AI	7B	18.1 GB estimated	+3.9 GB	✓ fits	weights 14.0 GB estimated KV cache 2.00 GB (heuristic) overhead 2.1 GB reserved 2 GB context 32,768 × batch 1
Pythia-Chat-Base-7BOpen weightsTogether AI	7B	18.1 GB estimated	+3.9 GB	✓ fits	weights 14.0 GB estimated KV cache 2.00 GB (heuristic) overhead 2.1 GB reserved 2 GB context 32,768 × batch 1
Qwen2.5-OmniOpen weightsQwen Team	7B	18.1 GB estimated	+3.9 GB	✓ fits	weights 14.0 GB estimated KV cache 2.00 GB (heuristic) overhead 2.1 GB reserved 2 GB context 32,768 × batch 1
pythia-6.9bOpen weightsEleutherAI	6.99B	18.1 GB estimated	+3.9 GB	✓ fits	weights 14.0 GB estimated KV cache 2.00 GB (heuristic) overhead 2.1 GB reserved 2 GB context 32,768 × batch 1
OLMoE-1B-7B-0125-InstructOpen weightsAllen Institute for AI	6.92B	17.9 GB estimated	+4.1 GB	✓ fits	weights 13.8 GB estimated KV cache 2.00 GB (heuristic) overhead 2.1 GB reserved 2 GB context 32,768 × batch 1
OLMoE-1B-7B-0924Open weightsAllen Institute for AI	6.92B	17.9 GB estimated	+4.1 GB	✓ fits	weights 13.8 GB estimated KV cache 2.00 GB (heuristic) overhead 2.1 GB reserved 2 GB context 32,768 × batch 1
deepseek-coder-7b-instruct-v1.5Open weightsDeepSeek	6.91B	17.9 GB estimated	+4.1 GB	✓ fits	weights 13.8 GB estimated KV cache 2.00 GB (heuristic) overhead 2.1 GB reserved 2 GB context 32,768 × batch 1
deepseek-coder-6.7b-instructOpen weightsDeepSeek	6.74B	17.5 GB estimated	+4.5 GB	✓ fits	weights 13.5 GB estimated KV cache 2.00 GB (heuristic) overhead 2.0 GB reserved 2 GB context 32,768 × batch 1
NousResearch/Llama-2-7b-chat-hfOpen weightsNous Research	6.74B	17.5 GB estimated	+4.5 GB	✓ fits	weights 13.5 GB estimated KV cache 2.00 GB (heuristic) overhead 2.0 GB reserved 2 GB context 32,768 × batch 1
Llama-2-7b-chat-hfrestricted-weightsMeta AI	6.74B	17.5 GB estimated	+4.5 GB	✓ fits	weights 13.5 GB estimated KV cache 2.00 GB (heuristic) overhead 2.0 GB reserved 2 GB context 32,768 × batch 1
Llama-2-7b-hfrestricted-weightsMeta AI	6.74B	17.5 GB estimated	+4.5 GB	✓ fits	weights 13.5 GB estimated KV cache 2.00 GB (heuristic) overhead 2.0 GB reserved 2 GB context 32,768 × batch 1
NousResearch/Llama-2-7b-hfOpen weightsNous Research	6.74B	17.5 GB estimated	+4.5 GB	✓ fits	weights 13.5 GB estimated KV cache 2.00 GB (heuristic) overhead 2.0 GB reserved 2 GB context 32,768 × batch 1
Nous-Hermes-llama-2-7bOpen weightsNous Research	6.74B	17.5 GB estimated	+4.5 GB	✓ fits	weights 13.5 GB estimated KV cache 2.00 GB (heuristic) overhead 2.0 GB reserved 2 GB context 32,768 × batch 1
evo-1-131k-baseOpen weightsTogether AI	6.45B	16.9 GB estimated	+5.2 GB	✓ fits	weights 12.9 GB estimated KV cache 2.00 GB (heuristic) overhead 1.9 GB reserved 2 GB context 32,768 × batch 1
evo-1-8k-baseOpen weightsTogether AI	6.45B	16.9 GB estimated	+5.2 GB	✓ fits	weights 12.9 GB estimated KV cache 2.00 GB (heuristic) overhead 1.9 GB reserved 2 GB context 32,768 × batch 1
chatglm3-6bOpen weightsZ.ai (Zhipu AI)	6.24B	16.4 GB estimated	+5.6 GB	✓ fits	weights 12.5 GB estimated KV cache 2.00 GB (heuristic) overhead 1.9 GB reserved 2 GB context 32,768 × batch 1
chatglm2-6bOpen weightsZ.ai (Zhipu AI)	6B	15.8 GB estimated	+6.2 GB	✓ fits	weights 12.0 GB estimated KV cache 2.00 GB (heuristic) overhead 1.8 GB reserved 2 GB context 32,768 × batch 1
GPT-JT-6B-v1Open weightsTogether AI	6B	15.8 GB estimated	+6.2 GB	✓ fits	weights 12.0 GB estimated KV cache 2.00 GB (heuristic) overhead 1.8 GB reserved 2 GB context 32,768 × batch 1
gpt-j-6bOpen weightsEleutherAI	6B	15.8 GB estimated	+6.2 GB	✓ fits	weights 12.0 GB estimated KV cache 2.00 GB (heuristic) overhead 1.8 GB reserved 2 GB context 32,768 × batch 1
gemma-4-E2B-itOpen weightsGoogle	5.12B	13.8 GB estimated	+8.2 GB	✓ fits	weights 10.3 GB estimated KV cache 2.00 GB (heuristic) overhead 1.5 GB reserved 2 GB context 32,768 × batch 1
Molmo2-4BOpen weightsAllen Institute for AI	4.85B	13.2 GB estimated	+8.8 GB	✓ fits	weights 9.7 GB estimated KV cache 2.00 GB (heuristic) overhead 1.5 GB reserved 2 GB context 32,768 × batch 1
Qianfan-OCROpen weightsBaidu	4.74B	12.9 GB estimated	+9.1 GB	✓ fits	weights 9.5 GB estimated KV cache 2.00 GB (heuristic) overhead 1.4 GB reserved 2 GB context 32,768 × batch 1
Voxtral-Mini-3B-2507Open weightsMistral AI	4.68B	12.8 GB estimated	+9.3 GB	✓ fits	weights 9.3 GB estimated KV cache 2.00 GB (heuristic) overhead 1.4 GB reserved 2 GB context 32,768 × batch 1
Qwen3.5-4BOpen weightsQwen	4.66B	12.7 GB estimated	+9.3 GB	✓ fits	weights 9.3 GB estimated KV cache 2.00 GB (heuristic) overhead 1.4 GB reserved 2 GB context 32,768 × batch 1
Qwen3-VL-4B-InstructOpen weightsQwen	4.44B	12.2 GB estimated	+9.8 GB	✓ fits	weights 8.9 GB estimated KV cache 2.00 GB (heuristic) overhead 1.3 GB reserved 2 GB context 32,768 × batch 1
Voxtral-Mini-4B-Realtime-2602Open weightsMistral AI	4.43B	12.2 GB estimated	+9.8 GB	✓ fits	weights 8.9 GB estimated KV cache 2.00 GB (heuristic) overhead 1.3 GB reserved 2 GB context 32,768 × batch 1
Gemma 3 4Brestricted-weightsGoogle	4.3B	11.9 GB estimated	+10.1 GB	✓ fits	weights 8.6 GB estimated KV cache 2.00 GB (heuristic) overhead 1.3 GB reserved 2 GB context 32,768 × batch 1
Ministral-3-3B-Reasoning-2512Open weightsMistral AI	4.25B	11.8 GB estimated	+10.2 GB	✓ fits	weights 8.5 GB estimated KV cache 2.00 GB (heuristic) overhead 1.3 GB reserved 2 GB context 32,768 × batch 1
Phi-3.5-vision-instructOpen weightsMicrosoft	4.15B	11.5 GB estimated	+10.5 GB	✓ fits	weights 8.3 GB estimated KV cache 2.00 GB (heuristic) overhead 1.2 GB reserved 2 GB context 32,768 × batch 1
instructblip-flan-t5-xlOpen weightsSalesforce	4.02B	11.3 GB estimated	+10.7 GB	✓ fits	weights 8.1 GB estimated KV cache 2.00 GB (heuristic) overhead 1.2 GB reserved 2 GB context 32,768 × batch 1
Qwen3-4BOpen weightsQwen	4.02B	11.3 GB estimated	+10.8 GB	✓ fits	weights 8.0 GB estimated KV cache 2.00 GB (heuristic) overhead 1.2 GB reserved 2 GB context 32,768 × batch 1
pplx-embed-context-v1-4bOpen weightsPerplexity AI	4.02B	11.3 GB estimated	+10.8 GB	✓ fits	weights 8.0 GB estimated KV cache 2.00 GB (heuristic) overhead 1.2 GB reserved 2 GB context 32,768 × batch 1
pplx-embed-v1-4bOpen weightsPerplexity AI	4.02B	11.3 GB estimated	+10.8 GB	✓ fits	weights 8.0 GB estimated KV cache 2.00 GB (heuristic) overhead 1.2 GB reserved 2 GB context 32,768 × batch 1
TRELLIS.2-4BOpen weightsMicrosoft	4B	11.2 GB estimated	+10.8 GB	✓ fits	weights 8.0 GB estimated KV cache 2.00 GB (heuristic) overhead 1.2 GB reserved 2 GB context 32,768 × batch 1
granite-vision-4.1-4bOpen weightsIBM	4B	11.2 GB estimated	+10.8 GB	✓ fits	weights 8.0 GB estimated KV cache 2.00 GB (heuristic) overhead 1.2 GB reserved 2 GB context 32,768 × batch 1
blip2-flan-t5-xlOpen weightsSalesforce	3.94B	11.1 GB estimated	+10.9 GB	✓ fits	weights 7.9 GB estimated KV cache 2.00 GB (heuristic) overhead 1.2 GB reserved 2 GB context 32,768 × batch 1
FLUX.2-klein-base-4BOpen weightsBlack Forest Labs	3.88B	10.9 GB estimated	+11.1 GB	✓ fits	weights 7.8 GB estimated KV cache 2.00 GB (heuristic) overhead 1.2 GB reserved 2 GB context 32,768 × batch 1
FLUX.2-klein-4BOpen weightsBlack Forest Labs	3.88B	10.9 GB estimated	+11.1 GB	✓ fits	weights 7.8 GB estimated KV cache 2.00 GB (heuristic) overhead 1.2 GB reserved 2 GB context 32,768 × batch 1
Cosmos3-EdgeOpen weightsNVIDIA	3.86B	10.9 GB estimated	+11.1 GB	✓ fits	weights 7.7 GB estimated KV cache 2.00 GB (heuristic) overhead 1.2 GB reserved 2 GB context 32,768 × batch 1
Ministral 3 3B 2512Open weightsMistral AI	3.85B	10.8 GB estimated	+11.2 GB	✓ fits	weights 7.7 GB estimated KV cache 2.00 GB (heuristic) overhead 1.1 GB reserved 2 GB context 32,768 × batch 1
CARE-XProprietaryMicrosoft	3.8B	10.7 GB estimated	+11.3 GB	✓ fits	weights 7.6 GB estimated KV cache 2.00 GB (heuristic) overhead 1.1 GB reserved 2 GB context 32,768 × batch 1

First 120 models (fitting first). The full list is in Run locally.

MethodEstimated, not measured: weights = parameters × bytes/param × 1.15 runtime overhead (or the observed artifact file size when one is recorded). /methodology

Relations

In the graph

Explore graph →

Manufactured by: Apple
Uses: Apple M5

Timeline

Events

Full timeline →

No events yet

Events are generated by the change engine when a material property, price or result changes.

Sources

Where these facts come from

Source documents
Source	Document	Type	Tier	Last observed	Snapshots
Apple — Mac tech specs	apple.com/macbook-air/specs	spec_page	T1· Official	11 h ago	1

Tier 1 = official/primary, 2 = quality secondary, 3 = community, 4 = unverified. Every snapshot is archived; see all sources and the methodology.