SCX (Global)

SCX directly confirmed zero data retention across its inference catalogue.

Data handling can vary by model and deployment. The privacy badge above shows the more conservative of the provider-wide fallback and the policies for routes on this page; each model page shows the policy for that specific route.

Models served by SCX (Global)

Model name

17 models

Model	Context	Input /M	Output /M	Cache read /M	Cache hit	TTFT	Throughput
DeepSeek V4.1 Flash FP8 deepseek/deepseek-v4.1-flash:scx	1.0M	$0.16/M	$0.63/M	$0.01/M	94.2%	6.9s	160.2 tps
DeepSeek V4.1 Flash (Thinking) FP8 deepseek/deepseek-v4.1-flash:thinking:scx	1.0M	$0.16/M	$0.63/M	$0.01/M	94.2%	6.9s	160.2 tps
Gemma 4 31B BF16 google/gemma-4-31b-it:scx	131k	$0.31/M	$0.96/M	N/A	0%	3.4s	321.8 tps
Gemma 4 31B (Thinking) BF16 google/gemma-4-31b-it:thinking:scx	131k	$0.31/M	$0.96/M	N/A	0%	3.4s	321.8 tps
GLM 5.2 BF16 z-ai/glm-5.2:scx	1.0M	$0.64/M	$2.23/M	$0.16/M	N/A	N/A	N/A
GLM 5.2 (Thinking) BF16 z-ai/glm-5.2:thinking:scx	1.0M	$0.64/M	$2.23/M	$0.16/M	49.6%	5.2s	59.8 tps
GLM 5.3 BF16 z-ai/glm-5.3:scx	1.0M	$0.97/M	$3.05/M	$0.18/M	65.5%	5.2s	44.7 tps
GLM 5.3 (Thinking) BF16 z-ai/glm-5.3:thinking:scx	1.0M	$0.97/M	$3.05/M	$0.18/M	65.5%	6s	35.2 tps
GLM 5.3 Flash FP8 z-ai/glm-5.3-flash:scx	1.0M	$0.09/M	$0.26/M	$0.03/M	91%	9.6s	64.7 tps
GPT OSS 120B FP8 openai/gpt-oss-120b:scx	131k	$0.18/M	$0.58/M	N/A	N/A	N/A	N/A
Kimi K2.7 Code FP8 moonshotai/kimi-k2.7-code:scx	262k	$0.77/M	$3.23/M	$0.15/M	N/A	N/A	N/A
Kimi K3 FP8 moonshotai/kimi-k3:scx	1.0M	$3.29/M	$16.46/M	$0.33/M	N/A	N/A	N/A
Llama 4 Maverick FP8 meta-llama/llama-4-maverick:scx	131k	$0.56/M	$1.70/M	N/A	N/A	N/A	N/A
MiniMax M2.7 FP8 minimax/minimax-m2.7:scx	192k	$0.50/M	$1.88/M	$0.05/M	N/A	N/A	N/A
Qwen3 32B BF16 qwen/qwen3-32b:scx	33k	$0.38/M	$0.91/M	N/A	N/A	N/A	N/A
Qwen3.8 Max qwen3.8-max:scx	1M	$1.62/M	$4.86/M	$0.21/M	N/A	N/A	N/A
Qwen3.8 Max Thinking qwen3.8-max:thinking:scx	1M	$1.62/M	$4.86/M	$0.21/M	N/A	N/A	N/A

DeepSeek V4.1 Flash

FP8

deepseek/deepseek-v4.1-flash:scx

Context

1.0M

Input /M

$0.16/M

Output /M

$0.63/M

Cache read /M

$0.01/M

Cache hit

94.2%

TTFT

6.9s

Throughput

160.2 tps

DeepSeek V4.1 Flash (Thinking)

FP8

deepseek/deepseek-v4.1-flash:thinking:scx

Context

1.0M

Input /M

$0.16/M

Output /M

$0.63/M

Cache read /M

$0.01/M

Cache hit

94.2%

TTFT

6.9s

Throughput

160.2 tps

Gemma 4 31B

BF16

google/gemma-4-31b-it:scx

Context

131k

Input /M

$0.31/M

Output /M

$0.96/M

Cache read /M

N/A

Cache hit

0%

TTFT

3.4s

Throughput

321.8 tps

Gemma 4 31B (Thinking)

BF16

google/gemma-4-31b-it:thinking:scx

Context

131k

Input /M

$0.31/M

Output /M

$0.96/M

Cache read /M

N/A

Cache hit

0%

TTFT

3.4s

Throughput

321.8 tps

GLM 5.2

BF16

z-ai/glm-5.2:scx

Context

1.0M

Input /M

$0.64/M

Output /M

$2.23/M

Cache read /M

$0.16/M

Cache hit

N/A

TTFT

N/A

Throughput

N/A

GLM 5.2 (Thinking)

BF16

z-ai/glm-5.2:thinking:scx

Context

1.0M

Input /M

$0.64/M

Output /M

$2.23/M

Cache read /M

$0.16/M

Cache hit

49.6%

TTFT

5.2s

Throughput

59.8 tps

GLM 5.3

BF16

z-ai/glm-5.3:scx

Context

1.0M

Input /M

$0.97/M

Output /M

$3.05/M

Cache read /M

$0.18/M

Cache hit

65.5%

TTFT

5.2s

Throughput

44.7 tps

GLM 5.3 (Thinking)

BF16

z-ai/glm-5.3:thinking:scx

Context

1.0M

Input /M

$0.97/M

Output /M

$3.05/M

Cache read /M

$0.18/M

Cache hit

65.5%

TTFT

6s

Throughput

35.2 tps

GLM 5.3 Flash

FP8

z-ai/glm-5.3-flash:scx

Context

1.0M

Input /M

$0.09/M

Output /M

$0.26/M

Cache read /M

$0.03/M

Cache hit

91%

TTFT

9.6s

Throughput

64.7 tps

GPT OSS 120B

FP8

openai/gpt-oss-120b:scx

Context

131k

Input /M

$0.18/M

Output /M

$0.58/M

Cache read /M

N/A

Cache hit

N/A

TTFT

N/A

Throughput

N/A

Kimi K2.7 Code

FP8

moonshotai/kimi-k2.7-code:scx

Context

262k

Input /M

$0.77/M

Output /M

$3.23/M

Cache read /M

$0.15/M

Cache hit

N/A

TTFT

N/A

Throughput

N/A

Kimi K3

FP8

moonshotai/kimi-k3:scx

Context

1.0M

Input /M

$3.29/M

Output /M

$16.46/M

Cache read /M

$0.33/M

Cache hit

N/A

TTFT

N/A

Throughput

N/A

Llama 4 Maverick

FP8

meta-llama/llama-4-maverick:scx

Context

131k

Input /M

$0.56/M

Output /M

$1.70/M

Cache read /M

N/A

Cache hit

N/A

TTFT

N/A

Throughput

N/A

MiniMax M2.7

FP8

minimax/minimax-m2.7:scx

Context

192k

Input /M

$0.50/M

Output /M

$1.88/M

Cache read /M

$0.05/M

Cache hit

N/A

TTFT

N/A

Throughput

N/A

Qwen3 32B

BF16

qwen/qwen3-32b:scx

Context

33k

Input /M

$0.38/M

Output /M

$0.91/M

Cache read /M

N/A

Cache hit

N/A

TTFT

N/A

Throughput

N/A

Qwen3.8 Max

qwen3.8-max:scx

Context

1M

Input /M

$1.62/M

Output /M

$4.86/M

Cache read /M

$0.21/M

Cache hit

N/A

TTFT

N/A

Throughput

N/A

Qwen3.8 Max Thinking

qwen3.8-max:thinking:scx

Context

1M

Input /M

$1.62/M

Output /M

$4.86/M

Cache read /M

$0.21/M

Cache hit

N/A

TTFT

N/A

Throughput

N/A

Prices are per million tokens. TTFT is the observed time to first token. Selectable routes include the applicable provider-selection markup; fixed routes show the standard model price. Availability and pricing refresh continuously; each model page shows the live provider comparison.

Using SCX (Global) via the API

Call a model served by SCX (Global) using its displayed model ID, including the provider suffix when selection is supported.

curl https://nano-gpt.com/api/v1/chat/completions \
  -H "Authorization: Bearer $NANOGPT_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "deepseek/deepseek-v4.1-flash:scx",
    "messages": [{"role": "user", "content": "Hello"}]
  }'

For models with provider selection, append :scx to the model ID, set "provider": "scx" in the request body, or send an X-Provider: scx header. Explicit provider selection adds a route-specific markup over that provider's base price; the prices in the table above already include it. Models without provider selection use the displayed model ID without a provider suffix or selection surcharge.

See the API documentation for provider preferences, price-aware routing, and error behavior.