Fireworks

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 Fireworks

Model name

19 models

Model	Context	Input /M	Output /M	Cache read /M	TTFT	Throughput
DeepSeek V4 Pro deepseek/deepseek-v4-pro:fireworks	1.0M	$1.83/M	$3.65/M	$0.15/M	N/A	N/A
DeepSeek V4.1 Flash deepseek/deepseek-v4.1-flash:fireworks	1.0M	$0.31/M	$1.26/M	$0.0063/M	1.3s	109 tps
DeepSeek V4.1 Flash (Thinking) deepseek/deepseek-v4.1-flash:thinking:fireworks	1.0M	$0.31/M	$1.26/M	$0.0063/M	1.3s	109 tps
Ember 1 fireworks/ember-1:fireworks	1.0M	$3.15/M	$15.75/M	$0.31/M	1.8s	56 tps
GLM 5.2 z-ai/glm-5.2:fireworks	1.0M	$1.47/M	$4.62/M	$0.27/M	0.9s	56.5 tps
GLM 5.2 Fast z-ai/glm-5.2:fireworks-fast	1.0M	$2.21/M	$6.93/M	$0.22/M	0.7s	108 tps
GLM 5.2 (Thinking) z-ai/glm-5.2:thinking:fireworks	1.0M	$1.47/M	$4.62/M	$0.27/M	0.9s	56.5 tps
GLM 5.2 (Thinking) Fast z-ai/glm-5.2:thinking:fireworks-fast	1.0M	$2.21/M	$6.93/M	$0.22/M	0.7s	108 tps
GLM 5.3 z-ai/glm-5.3:fireworks	1.0M	$1.47/M	$4.62/M	$0.27/M	1.9s	48 tps
GLM 5.3 (Thinking) z-ai/glm-5.3:thinking:fireworks	1.0M	$1.47/M	$4.62/M	$0.27/M	1.9s	48 tps
GLM 5.3 Flash z-ai/glm-5.3-flash:fireworks	1.0M	$0.16/M	$0.53/M	$0.03/M	1.1s	94 tps
Inkling thinkingmachines/inkling:fireworks	1.0M	$1.05/M	$4.25/M	$0.18/M	N/A	N/A
Inkling Thinking thinkingmachines/inkling:thinking:fireworks	1.0M	$1.05/M	$4.25/M	$0.18/M	N/A	N/A
Kimi K3 moonshotai/kimi-k3:fireworks	1.0M	$3.15/M	$15.75/M	$0.31/M	0.8s	41 tps
Kimi K3 Fast moonshotai/kimi-k3:fireworks-fast	1.0M	$4.73/M	$23.63/M	$0.47/M	1.3s	68 tps
Nvidia Nemotron 3.5 Lightning nvidia/nemotron-3.5-lightning:fireworks	262k	$0.05/M	$0.21/M	$0.01/M	N/A	N/A
Nvidia Nemotron 3.5 Lightning Thinking nvidia/nemotron-3.5-lightning:thinking:fireworks	262k	$0.05/M	$0.21/M	$0.01/M	N/A	N/A
Qwen3.8 Max qwen3.8-max:fireworks	262k	$2.10/M	$6.30/M	$0.26/M	N/A	N/A
Qwen3.8 Max Thinking qwen3.8-max:thinking:fireworks	262k	$2.10/M	$6.30/M	$0.26/M	N/A	N/A

DeepSeek V4 Pro

deepseek/deepseek-v4-pro:fireworks

Context

1.0M

Input /M

$1.83/M

Output /M

$3.65/M

Cache read /M

$0.15/M

TTFT

N/A

Throughput

N/A

DeepSeek V4.1 Flash

deepseek/deepseek-v4.1-flash:fireworks

Context

1.0M

Input /M

$0.31/M

Output /M

$1.26/M

Cache read /M

$0.0063/M

TTFT

1.3s

Throughput

109 tps

DeepSeek V4.1 Flash (Thinking)

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

Context

1.0M

Input /M

$0.31/M

Output /M

$1.26/M

Cache read /M

$0.0063/M

TTFT

1.3s

Throughput

109 tps

Ember 1

fireworks/ember-1:fireworks

Context

1.0M

Input /M

$3.15/M

Output /M

$15.75/M

Cache read /M

$0.31/M

TTFT

1.8s

Throughput

56 tps

GLM 5.2

z-ai/glm-5.2:fireworks

Context

1.0M

Input /M

$1.47/M

Output /M

$4.62/M

Cache read /M

$0.27/M

TTFT

0.9s

Throughput

56.5 tps

GLM 5.2

Fast

z-ai/glm-5.2:fireworks-fast

Context

1.0M

Input /M

$2.21/M

Output /M

$6.93/M

Cache read /M

$0.22/M

TTFT

0.7s

Throughput

108 tps

GLM 5.2 (Thinking)

z-ai/glm-5.2:thinking:fireworks

Context

1.0M

Input /M

$1.47/M

Output /M

$4.62/M

Cache read /M

$0.27/M

TTFT

0.9s

Throughput

56.5 tps

GLM 5.2 (Thinking)

Fast

z-ai/glm-5.2:thinking:fireworks-fast

Context

1.0M

Input /M

$2.21/M

Output /M

$6.93/M

Cache read /M

$0.22/M

TTFT

0.7s

Throughput

108 tps

GLM 5.3

z-ai/glm-5.3:fireworks

Context

1.0M

Input /M

$1.47/M

Output /M

$4.62/M

Cache read /M

$0.27/M

TTFT

1.9s

Throughput

48 tps

GLM 5.3 (Thinking)

z-ai/glm-5.3:thinking:fireworks

Context

1.0M

Input /M

$1.47/M

Output /M

$4.62/M

Cache read /M

$0.27/M

TTFT

1.9s

Throughput

48 tps

GLM 5.3 Flash

z-ai/glm-5.3-flash:fireworks

Context

1.0M

Input /M

$0.16/M

Output /M

$0.53/M

Cache read /M

$0.03/M

TTFT

1.1s

Throughput

94 tps

Inkling

thinkingmachines/inkling:fireworks

Context

1.0M

Input /M

$1.05/M

Output /M

$4.25/M

Cache read /M

$0.18/M

TTFT

N/A

Throughput

N/A

Inkling Thinking

thinkingmachines/inkling:thinking:fireworks

Context

1.0M

Input /M

$1.05/M

Output /M

$4.25/M

Cache read /M

$0.18/M

TTFT

N/A

Throughput

N/A

Kimi K3

moonshotai/kimi-k3:fireworks

Context

1.0M

Input /M

$3.15/M

Output /M

$15.75/M

Cache read /M

$0.31/M

TTFT

0.8s

Throughput

41 tps

Kimi K3

Fast

moonshotai/kimi-k3:fireworks-fast

Context

1.0M

Input /M

$4.73/M

Output /M

$23.63/M

Cache read /M

$0.47/M

TTFT

1.3s

Throughput

68 tps

Nvidia Nemotron 3.5 Lightning

nvidia/nemotron-3.5-lightning:fireworks

Context

262k

Input /M

$0.05/M

Output /M

$0.21/M

Cache read /M

$0.01/M

TTFT

N/A

Throughput

N/A

Nvidia Nemotron 3.5 Lightning Thinking

nvidia/nemotron-3.5-lightning:thinking:fireworks

Context

262k

Input /M

$0.05/M

Output /M

$0.21/M

Cache read /M

$0.01/M

TTFT

N/A

Throughput

N/A

Qwen3.8 Max

qwen3.8-max:fireworks

Context

262k

Input /M

$2.10/M

Output /M

$6.30/M

Cache read /M

$0.26/M

TTFT

N/A

Throughput

N/A

Qwen3.8 Max Thinking

qwen3.8-max:thinking:fireworks

Context

262k

Input /M

$2.10/M

Output /M

$6.30/M

Cache read /M

$0.26/M

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 Fireworks via the API

Call a model served by Fireworks 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-pro:fireworks",
    "messages": [{"role": "user", "content": "Hello"}]
  }'

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

Fireworks offers multiple service lanes, each pinned with its own id: fireworks, fireworks-fast (Fast). The table shows which lane serves each row.

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