FitMyLLM
QUALCOMM· ADRENO 800

Qualcomm Adreno 830

Running LLMs on the Qualcomm Adreno 830 — the long read: which models fit at which quantisation, and the settings worth changing. · Or what a budget buys

VRAM
16 GB
MID-RANGE
BANDWIDTH
77
GB/S
MODELS Q4
262/449
58%
7B Q4 SPEED
~10
USABLE
▸ MODEL COVERAGE @ Q458% OF ALL
▸ ESTIMATED SPEED· BY MODEL SIZE @ Q4

Average speeds at Q4 quantization. Actual performance varies by model architecture and context length.

3B
~23
TOK/S
7B
~10
TOK/S
14B
~5
TOK/S
32B
18.0GB NEEDED
70B
39.4GB NEEDED
▸ MEASURED RIG REPORTS

We rent the machine and time every model on it: decode, VRAM peak, concurrency, watts and cost per million tokens. Yours may already be one of them — and two are free to read in full.

SEE THE REPORTS
▸ SPECIFICATIONS
VRAM
16 GB
BANDWIDTH
77 GB/s
FP16 COMPUTE
4.6 TFLOPS
TDP
12W
MEMORY
Shared
ARCHITECTURE
Adreno 800
62
FAST MODELS · >30 TOK/S
Real-time chat speed
142
USABLE · >10 TOK/S
Comfortable for all tasks
262
TOTAL COMPATIBLE
Fit in VRAM at Q4
▸ DON’T WANT TO BUY?

Test Qualcomm Adreno 830 (or anything bigger) without committing. Pay by the second, cancel anytime.

Spin up in ~60s. Pay by the second. Cancel anytime.

Some links are affiliate links — we may earn a small commission at no extra cost to you. This helps keep FitMyLLM free and independent.

▸ COMPATIBLE MODELS· 262
S
Falcon-H1R Tiny 90M0.09B
FALCON·256K CTX· CHAT· REASONING
760
TOK/S · 3% VRAM
S
SmolLM2 135M0.135B
SMOLLM·2K CTX· CHAT
507
TOK/S · 4% VRAM
S
nomic-embed-text-v1.5 100M0.14B
EMBEDDING·8K CTX· CHAT
489
TOK/S · 4% VRAM
S
GPT-2 124M0.14B
GPT2·1K CTX· CHAT
489
TOK/S · 4% VRAM
S
SmolVLM 256M0.256B
SMOLLM·8K CTX· CHAT· VISION
267
TOK/S · 4% VRAM
S
Gemma 3 270M0.27B
GEMMA·32K CTX· CHAT
253
TOK/S · 4% VRAM
S
Snowflake Arctic Embed M v2.00.305B
EMBEDDING·8K CTX· EMBEDDING· MULTILINGUAL
224
TOK/S · 4% VRAM
S
bge-large-en-v1.5 335M0.335B
EMBEDDING·1K CTX· CHAT
204
TOK/S · 4% VRAM
S
mxbai-embed-large-v10.335B
EMBEDDING·1K CTX· EMBEDDING
204
TOK/S · 4% VRAM
S
Snowflake Arctic Embed L0.335B
EMBEDDING·1K CTX· EMBEDDING
204
TOK/S · 4% VRAM
S
LFM2 350M0.35B
LFM·125K CTX· CHAT· TOOL_USE
196
TOK/S · 4% VRAM
S
SmolLM2 360M0.36B
SMOLLM·8K CTX· CHAT
190
TOK/S · 4% VRAM
S
GPT-2 Medium 345M0.38B
GPT2·1K CTX· CHAT
180
TOK/S · 5% VRAM
S
Qwen 2.5 0.5B0.5B
QWEN·32K CTX· CHAT
137
TOK/S · 5% VRAM
S
SmolVLM 500M0.5B
SMOLLM·8K CTX· CHAT· VISION
137
TOK/S · 5% VRAM
S
Falcon-H1 0.5B0.52B
FALCON·128K CTX· CHAT
132
TOK/S · 5% VRAM
S
BGE-M30.568B
EMBEDDING·8K CTX· EMBEDDING
121
TOK/S · 5% VRAM
S
Snowflake Arctic Embed L v2.00.568B
EMBEDDING·8K CTX· EMBEDDING· MULTILINGUAL
121
TOK/S · 5% VRAM
S
Qwen3-Embedding 0.6B0.6B
EMBEDDING·32K CTX· EMBEDDING· MULTILINGUAL
114
TOK/S · 5% VRAM
S
Falcon-H1R Tiny 0.6B0.6B
FALCON·32K CTX· CHAT· REASONING
114
TOK/S · 5% VRAM
S
Falcon Perception 0.6B0.6B
FALCON·4K CTX· VISION
114
TOK/S · 5% VRAM
S
Qwen 1.5 0.5B0.62B
QWEN·32K CTX· CHAT
110
TOK/S · 5% VRAM
S
LFM2 700M0.74B
LFM·125K CTX· CHAT· TOOL_USE
92
TOK/S · 6% VRAM
S
Qwen3 0.6B0.75B
QWEN·32K CTX· CHAT· REASONING
91
TOK/S · 6% VRAM
S
GPT-2 Large 774M0.81B
GPT2·1K CTX· CHAT
84
TOK/S · 6% VRAM
S
Qwen 3.5 0.8B0.87B
QWEN·256K CTX· CHAT· CODING· MULTILINGUAL
79
TOK/S · 6% VRAM
S
InternVL3 1B0.94B
OTHER·32K CTX· CHAT· VISION
73
TOK/S · 7% VRAM
S
MiniCPM5 1B1.08B
MINICPM·128K CTX· CHAT· REASONING· MULTILINGUAL
63
TOK/S · 7% VRAM
S
TinyLlama 1.1B1.1B
LLAMA·2K CTX· CHAT
62
TOK/S · 7% VRAM
A
LFM2.5-1.2B-Thinking1.2B
LFM·122K CTX· CHAT· REASONING· TOOL_USE
57
TOK/S · 8% VRAM
A
Llama-3.2-1B1.2B
LLAMA·4K CTX· CHAT
57
TOK/S · 8% VRAM
A
LFM2 1.2B1.2B
LFM·125K CTX· CHAT· TOOL_USE· MULTILINGUAL
57
TOK/S · 8% VRAM
A
Zamba2 1.2B1.2B
OTHER·4K CTX· CHAT
57
TOK/S · 8% VRAM
A
EXAONE-4.0-1.2B1.3B
EXAONE·64K CTX· CHAT
53
TOK/S · 8% VRAM
A
OPT 1.3B1.3B
OPT·2K CTX· CHAT
53
TOK/S · 8% VRAM
A
MiniCPM-V 4.61.3B
OTHER·256K CTX· CHAT· VISION
53
TOK/S · 8% VRAM
A
DeepSeek Coder 1.3B1.35B
DEEPSEEK·16K CTX· CODING
51
TOK/S · 8% VRAM
A
Phi-1 1.3B1.42B
PHI·2K CTX· CODING
48
TOK/S · 8% VRAM
A
Phi-1.5 1.3B1.42B
PHI·2K CTX· CHAT· CODING
48
TOK/S · 8% VRAM
A
LFM2 8B A1B8.3BMoE
LFM·125K CTX· CHAT· TOOL_USE· MULTILINGUAL
46
TOK/S · 35% VRAM
▸ NEXT STEP

Get personalized recommendations.

See ranked models with benchmark scores, run commands, and precise speed estimates for your Qualcomm Adreno 830.

WHAT THIS CARD IS WORTH

Qualcomm Adreno 830 holds 262 of the models in our catalogue and is, in practice, a IQ4_XS card — the largest it takes is ERNIE 4.5 21B A3B at IQ4_XS.

TOKENS/SEC PER $100
8B at Q4_K_M, so cards compare like for like
VRAM PER $100
what memory costs on this card
THE BIGGEST IT TAKES
ERNIE 4.5 21B A3B21.95B · IQ4_XS13.9 GB35 tok/sEST
GPT-OSS 20B21B · Q4_K_S13.8 GB27 tok/sEST
Reka Flash 321B · Q4_K_S14.2 GB5 tok/sEST
Reka Flash 3.121B · Q4_K_S14.2 GB5 tok/sEST
InternLM2 20B19.8B · Q4_K_M14.3 GB5 tok/sEST
InternLM2.5 20B19.8B · Q4_K_M14.3 GB5 tok/sEST
Ling-lite 16.8B16.8B · Q5_K_M13.6 GB34 tok/sEST
DeepSeek V2 Lite 16B16B · Q5_K_M12.9 GB34 tok/sEST
AGAINST CARDS OF SIMILAR MEMORY
M1 Max (32GB)21 GB$1,4993.0 tok/s per $100
M2 Max (32GB)21 GB$1,7992.5 tok/s per $100
M2 Pro (32GB)21 GB$1,4991.9 tok/s per $100
M4 (32GB)21 GB$1,1991.7 tok/s per $100
Qualcomm Adreno 83016 GB

Neighbours in memory rather than in price: memory decides whether a card can do the job at all, so two cards of the same size at different prices is the comparison you are making. Ordered by memory, then bandwidth — not by the value column, which is worked out from bandwidth and price alone and therefore rewards a cheap card whatever its software stack does to that bandwidth in practice. Read it as one input, not as a ranking.

▸ DEVICE UNDER TEST

Qualcomm Adreno 830 — 16 GB VRAM.

QUALCOMM ADRENO 830 SPEC
BRAND
Apple
VRAM
16 GB Shared
BANDWIDTH
77 GB/s
FP16 COMPUTE
4.6 TFLOPS
TDP
12 W
ARCHITECTURE
Adreno 800
▸ AI CAPABILITY
262/ 449 models @ Q4

With 16 GB VRAM and 77 GB/s bandwidth, this GPU handles models up to 19.8B parameters.

Speed ≈ bandwidth / model_size × efficiency. A 7B model at Q4 runs at ~9 tok/s.

RENT IT — LIVE PRICES
Checking what the clouds are charging…
§ 01TOP MODELS FOR QUALCOMM ADRENO 830
262 FIT · SHOWING 20
MODELSIZEVRAM Q4TOK/SAVG
InternLM2 20B19.8B12.6 GB345.1
InternLM2.5 20B19.8B12.6 GB350.9
Ling-lite 16.8B16.8B10.8 GB29
DeepSeek V2 Lite 16B16B10.3 GB2938.0
StarCoder2 15B15.96B10.2 GB426.5
DeepSeek-Coder-V2-Lite 15.7B15.7B10.1 GB2943.0
DeepSeek-VL2 Small 16B15.7B10.1 GB2943.1
StarCoder 15B15.5B10.0 GB421.0
InternVL3 14B15.12B9.7 GB538.1
Phi-4-reasoning-vision 15B15B9.7 GB542.8
DeepSeek R1 Distill Qwen 14B14.8B9.5 GB543.9
DeepCoder 14B14.8B9.5 GB538.7
Qwen2.5-Coder-14B14.8B9.5 GB541.3
Qwen2.5-14B14.8B9.5 GB541.3
Qwen3 14B14.8B9.5 GB545.7
phi-4 14B14.66B9.4 GB533.7
Phi-4-reasoning 14B14.66B9.4 GB533.7
Phi-4-reasoning-plus 14B14.66B9.4 GB575.5
Ministral 3 14B14B9.0 GB525.9
Phi-3-medium-14b14B9.0 GB533.7