Reka

Models served by Reka

Model name

15 models

Model	Context	Input /M	Output /M	Cache read /M	Cache hit	TTFT	Throughput
DeepSeek V4 Flash 0731 Unverified deepseek/deepseek-v4-flash-0731:reka	262k	$0.0515/M	$0.105/M	$0.0399/M	15.8%	1.3s	89.3 tps
DeepSeek V4 Flash 0731 (Thinking) Unverified deepseek/deepseek-v4-flash-0731:thinking:reka	262k	$0.0515/M	$0.105/M	$0.0399/M	45.6%	1.6s	75.2 tps
DeepSeek V4.1 Flash Unverified deepseek/deepseek-v4.1-flash:reka	1.0M	$0.031/M	$0.126/M	$0.00735/M	8.5%	3.3s	61.8 tps
DeepSeek V4.1 Flash (Thinking) Unverified deepseek/deepseek-v4.1-flash:thinking:reka	1.0M	$0.031/M	$0.126/M	$0.00735/M	88.4%	1.4s	139.2 tps
DeepSeek V4.1 Flash 1017 Unverified deepseek/deepseek-v4.1-flash-1017:reka	1.0M	$0.0546/M	$0.378/M	$0.0047/M	0%	1.9s	121.5 tps
DeepSeek V4.1 Flash 1017 (Thinking) Unverified deepseek/deepseek-v4.1-flash-1017:thinking:reka	1.0M	$0.0546/M	$0.378/M	$0.0047/M	86.5%	2.2s	356.9 tps
GLM 5.3 Unverified z-ai/glm-5.3:reka	262k	$0.21/M	$0.877/M	$0.142/M	7.4%	2.6s	40.9 tps
GLM 5.3 (Thinking) Unverified z-ai/glm-5.3:thinking:reka	262k	$0.21/M	$0.877/M	$0.142/M	0%	10.1s	38 tps
GLM 5.3 Flash Unverified z-ai/glm-5.3-flash:reka	1.0M	$0.0462/M	$0.315/M	$0.0168/M	93.3%	2.3s	273.6 tps
Qwen 3.6 27B Unverified alibaba/qwen3.6-27b:reka	262k	$0.157/M	$1.47/M	$0.0284/M	N/A	N/A	N/A
Qwen 3.6 27B Thinking Unverified alibaba/qwen3.6-27b:thinking:reka	262k	$0.157/M	$1.47/M	$0.0284/M	N/A	N/A	N/A
Qwen3.8 27B Unverified qwen3.8-27b:reka	262k	$0.0367/M	$0.331/M	$0.0189/M	N/A	N/A	N/A
Qwen3.8 27B Thinking Unverified qwen3.8-27b:thinking:reka	262k	$0.0367/M	$0.331/M	$0.0189/M	N/A	N/A	N/A
Reka Edge 2603 Unverified reka/reka-edge-2603:reka	16k	$0.105/M	$0.105/M	Free	N/A	N/A	N/A
Reka Flash 3 Unverified reka/reka-flash-3:reka	66k	$0.105/M	$0.21/M	Free	N/A	0.6s	96.2 tps

DeepSeek V4 Flash 0731

Unverified

deepseek/deepseek-v4-flash-0731:reka

Context

262k

Input /M

$0.0515/M

Output /M

$0.105/M

Cache read /M

$0.0399/M

Cache hit

15.8%

TTFT

1.3s

Throughput

89.3 tps

DeepSeek V4 Flash 0731 (Thinking)

Unverified

deepseek/deepseek-v4-flash-0731:thinking:reka

Context

262k

Input /M

$0.0515/M

Output /M

$0.105/M

Cache read /M

$0.0399/M

Cache hit

45.6%

TTFT

1.6s

Throughput

75.2 tps

DeepSeek V4.1 Flash

Unverified

deepseek/deepseek-v4.1-flash:reka

Context

1.0M

Input /M

$0.031/M

Output /M

$0.126/M

Cache read /M

$0.00735/M

Cache hit

8.5%

TTFT

3.3s

Throughput

61.8 tps

DeepSeek V4.1 Flash (Thinking)

Unverified

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

Context

1.0M

Input /M

$0.031/M

Output /M

$0.126/M

Cache read /M

$0.00735/M

Cache hit

88.4%

TTFT

1.4s

Throughput

139.2 tps

DeepSeek V4.1 Flash 1017

Unverified

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

Context

1.0M

Input /M

$0.0546/M

Output /M

$0.378/M

Cache read /M

$0.0047/M

Cache hit

0%

TTFT

1.9s

Throughput

121.5 tps

DeepSeek V4.1 Flash 1017 (Thinking)

Unverified

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

Context

1.0M

Input /M

$0.0546/M

Output /M

$0.378/M

Cache read /M

$0.0047/M

Cache hit

86.5%

TTFT

2.2s

Throughput

356.9 tps

GLM 5.3

Unverified

z-ai/glm-5.3:reka

Context

262k

Input /M

$0.21/M

Output /M

$0.877/M

Cache read /M

$0.142/M

Cache hit

7.4%

TTFT

2.6s

Throughput

40.9 tps

GLM 5.3 (Thinking)

Unverified

z-ai/glm-5.3:thinking:reka

Context

262k

Input /M

$0.21/M

Output /M

$0.877/M

Cache read /M

$0.142/M

Cache hit

0%

TTFT

10.1s

Throughput

38 tps

GLM 5.3 Flash

Unverified

z-ai/glm-5.3-flash:reka

Context

1.0M

Input /M

$0.0462/M

Output /M

$0.315/M

Cache read /M

$0.0168/M

Cache hit

93.3%

TTFT

2.3s

Throughput

273.6 tps

Qwen 3.6 27B

Unverified

alibaba/qwen3.6-27b:reka

Context

262k

Input /M

$0.157/M

Output /M

$1.47/M

Cache read /M

$0.0284/M

Cache hit

N/A

TTFT

N/A

Throughput

N/A

Qwen 3.6 27B Thinking

Unverified

alibaba/qwen3.6-27b:thinking:reka

Context

262k

Input /M

$0.157/M

Output /M

$1.47/M

Cache read /M

$0.0284/M

Cache hit

N/A

TTFT

N/A

Throughput

N/A

Qwen3.8 27B

Unverified

qwen3.8-27b:reka

Context

262k

Input /M

$0.0367/M

Output /M

$0.331/M

Cache read /M

$0.0189/M

Cache hit

N/A

TTFT

N/A

Throughput

N/A

Qwen3.8 27B Thinking

Unverified

qwen3.8-27b:thinking:reka

Context

262k

Input /M

$0.0367/M

Output /M

$0.331/M

Cache read /M

$0.0189/M

Cache hit

N/A

TTFT

N/A

Throughput

N/A

Reka Edge 2603

Unverified

reka/reka-edge-2603:reka

Context

16k

Input /M

$0.105/M

Output /M

$0.105/M

Cache read /M

Free

Cache hit

N/A

TTFT

N/A

Throughput

N/A

Reka Flash 3

Unverified

reka/reka-flash-3:reka

Context

66k

Input /M

$0.105/M

Output /M

$0.21/M

Cache read /M

Free

Cache hit

N/A

TTFT

0.6s

Throughput

96.2 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 Reka via the API

Call a model served by Reka 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-flash-0731:reka",
    "messages": [{"role": "user", "content": "Hello"}]
  }'

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