ModelRun

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 ModelRun

Model name

10 models

Model	Context	Input /M	Output /M	Cache read /M	Cache hit	TTFT	Throughput
DeepSeek V4.1 Flash 1017 FP4 deepseek/deepseek-v4.1-flash-1017:modelrun	1.0M	$0.178/M	$0.704/M	$0.0063/M	N/A	0.4s	124 tps
DeepSeek V4.1 Flash 1017 (Thinking) FP4 deepseek/deepseek-v4.1-flash-1017:thinking:modelrun	1.0M	$0.178/M	$0.704/M	$0.0063/M	N/A	0.4s	124 tps
Gemma 4 31B FP4 google/gemma-4-31b-it:modelrun	262k	$0.787/M	$1.05/M	$0.21/M	5%	0.2s	166 tps
Gemma 4 31B (Thinking) FP4 google/gemma-4-31b-it:thinking:modelrun	262k	$0.787/M	$1.05/M	$0.21/M	3.7%	0.2s	166 tps
GLM 5.3 FP4 z-ai/glm-5.3:modelrun	1.0M	$2.21/M	$6.93/M	$0.42/M	76.4%	1s	67 tps
GLM 5.3 (Thinking) FP4 z-ai/glm-5.3:thinking:modelrun	1.0M	$2.21/M	$6.93/M	$0.42/M	76.4%	1s	67 tps
MiniMax M3 FP8 minimax/minimax-m3:modelrun	1.0M	$0.315/M	$1.36/M	$0.0525/M	N/A	0.5s	82 tps
MiniMax M3 Thinking FP8 minimax/minimax-m3:thinking:modelrun	1.0M	$0.315/M	$1.36/M	$0.0525/M	N/A	0.5s	82 tps
Qwen3.8 27B FP4 qwen3.8-27b:modelrun	262k	$0.21/M	$2.63/M	$0.0945/M	N/A	0.2s	171 tps
Qwen3.8 27B Thinking FP4 qwen3.8-27b:thinking:modelrun	262k	$0.21/M	$2.63/M	$0.0945/M	N/A	0.2s	171 tps

DeepSeek V4.1 Flash 1017

FP4

deepseek/deepseek-v4.1-flash-1017:modelrun

Context

1.0M

Input /M

$0.178/M

Output /M

$0.704/M

Cache read /M

$0.0063/M

Cache hit

N/A

TTFT

0.4s

Throughput

124 tps

DeepSeek V4.1 Flash 1017 (Thinking)

FP4

deepseek/deepseek-v4.1-flash-1017:thinking:modelrun

Context

1.0M

Input /M

$0.178/M

Output /M

$0.704/M

Cache read /M

$0.0063/M

Cache hit

N/A

TTFT

0.4s

Throughput

124 tps

Gemma 4 31B

FP4

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

Context

262k

Input /M

$0.787/M

Output /M

$1.05/M

Cache read /M

$0.21/M

Cache hit

5%

TTFT

0.2s

Throughput

166 tps

Gemma 4 31B (Thinking)

FP4

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

Context

262k

Input /M

$0.787/M

Output /M

$1.05/M

Cache read /M

$0.21/M

Cache hit

3.7%

TTFT

0.2s

Throughput

166 tps

GLM 5.3

FP4

z-ai/glm-5.3:modelrun

Context

1.0M

Input /M

$2.21/M

Output /M

$6.93/M

Cache read /M

$0.42/M

Cache hit

76.4%

TTFT

1s

Throughput

67 tps

GLM 5.3 (Thinking)

FP4

z-ai/glm-5.3:thinking:modelrun

Context

1.0M

Input /M

$2.21/M

Output /M

$6.93/M

Cache read /M

$0.42/M

Cache hit

76.4%

TTFT

1s

Throughput

67 tps

MiniMax M3

FP8

minimax/minimax-m3:modelrun

Context

1.0M

Input /M

$0.315/M

Output /M

$1.36/M

Cache read /M

$0.0525/M

Cache hit

N/A

TTFT

0.5s

Throughput

82 tps

MiniMax M3 Thinking

FP8

minimax/minimax-m3:thinking:modelrun

Context

1.0M

Input /M

$0.315/M

Output /M

$1.36/M

Cache read /M

$0.0525/M

Cache hit

N/A

TTFT

0.5s

Throughput

82 tps

Qwen3.8 27B

FP4

qwen3.8-27b:modelrun

Context

262k

Input /M

$0.21/M

Output /M

$2.63/M

Cache read /M

$0.0945/M

Cache hit

N/A

TTFT

0.2s

Throughput

171 tps

Qwen3.8 27B Thinking

FP4

qwen3.8-27b:thinking:modelrun

Context

262k

Input /M

$0.21/M

Output /M

$2.63/M

Cache read /M

$0.0945/M

Cache hit

N/A

TTFT

0.2s

Throughput

171 tps

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 ModelRun via the API

Call a model served by ModelRun 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-1017:modelrun",
    "messages": [{"role": "user", "content": "Hello"}]
  }'

For models with provider selection, append :modelrun to the model ID, set "provider": "modelrun" in the request body, or send an X-Provider: modelrun 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.