Kitani

GLB

39 models available

NanoGPT requires zero data retention for Kitani routes serving Qwen 3.8 27B, GLM 5.3, Kimi K3, DeepSeek V4.1 Flash, GLM 5.3 Flash Uncensored, MiMo V2.6 Flash RL, and MiMo V2.6 Pro RL, including applicable thinking variants. Prompts and outputs are not retained or used for training; usage and billing metadata are retained. Kitani documents a 503 response when no qualified ZDR deployment is available. Its catalog reports encrypted inference transport (TLS) for the reviewed DeepSeek V4.1 Flash, uncensored, and MiMo RL routes. This is not a provider-wide ZDR commitment.

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 Kitani

39 models

Model	Context	Input /M	Output /M	Cache read /M	Cache hit	TTFT	Throughput
DeepSeek V4.1 Flash	262k	$0.05/M	$0.11/M	$0.0008/M	N/A	N/A	N/A
DeepSeek V4.1 Flash (Thinking)	262k	$0.05/M	$0.11/M	$0.0008/M	N/A	N/A	N/A
Gemma 3 12B IT INT4	8k	$0.26/M	$0.26/M	$0.13/M	N/A	N/A	N/A
Gemma 3 4B IT INT4	8k	$0.19/M	$0.19/M	$0.09/M	N/A	N/A	N/A
Gemma 4 12B Instruct INT4	33k	$0.05/M	$0.24/M	$0.02/M	N/A	N/A	N/A
Gemma 4 31B Gembrain uncensored heretic INT4	66k	$0.31/M	$0.31/M	$0.05/M	N/A	N/A	N/A
Gemma 4 31B Gembrain X Core INT4	66k	$0.31/M	$0.31/M	$0.05/M	N/A	N/A	N/A
Gemma 4 31B MeroMero v2 INT4	66k	$0.09/M	$0.46/M	$0.05/M	48.1%	2.5s	12.5 tps
Gemma 4 31B Split-Untied	262k	$0.11/M	$0.31/M	N/A	0%	1.3s	39.2 tps
Gemma 4 Dark Thoughts V2 31B INT4	66k	$0.11/M	$0.47/M	$0.05/M	N/A	N/A	N/A
Gemma 4 Giftige-Blume V2 31B INT4	66k	$0.11/M	$0.47/M	$0.05/M	N/A	N/A	N/A
Gemma 4 Gutenberg 31B INT4	66k	$0.11/M	$0.47/M	$0.05/M	N/A	N/A	N/A
Gemsicle INT4	66k	$0.09/M	$0.46/M	$0.05/M	N/A	N/A	N/A
GLM 5.3	1M	$0.58/M	$2.10/M	$0.12/M	49.5%	2.8s	148.5 tps
GLM 5.3 (Thinking)	1M	$0.58/M	$2.10/M	$0.12/M	49.6%	2.3s	111.6 tps
GLM 5.3 Flash Uncensored NVFP4	524k	$0.21/M	$0.84/M	$0.01/M	3.9%	17.6s	24.4 tps
GrayLine Qwen3 8B	33k	$0.29/M	$0.29/M	$0.15/M	N/A	N/A	N/A
LFM2.5 2.6B	8k	$0.09/M	$0.19/M	$0.05/M	N/A	N/A	N/A
Llama 3.1 8B INT4	8k	$0.05/M	$0.08/M	$0.03/M	N/A	N/A	N/A
Llama 3.2 3B Instruct INT4	8k	$0.03/M	$0.05/M	$0.01/M	N/A	N/A	N/A
MiMo V2.6 Pro	1.0M	$0.54/M	$1.10/M	$0.0045/M	32.3%	2.8s	23.4 tps
Ministral 3B INT4	8k	$0.09/M	$0.09/M	$0.05/M	N/A	N/A	N/A
Ministral 8B INT4	8k	$0.14/M	$0.14/M	$0.07/M	N/A	N/A	N/A
Mistral Nemo INT4	8k	$0.09/M	$0.11/M	$0.05/M	N/A	N/A	N/A
Mistral Small 3 24B (2501) INT4	8k	$0.11/M	$0.34/M	$0.05/M	N/A	N/A	N/A
Mistral Small 3.2 24B (2506) INT4	8k	$0.19/M	$0.40/M	$0.09/M	N/A	N/A	N/A
Neural Daredevil 8B abliterated INT4	8k	$0.44/M	$0.44/M	$0.22/M	N/A	N/A	N/A
Phi 4 Mini INT4	8k	$0.16/M	$0.69/M	$0.08/M	N/A	N/A	N/A
Qwen 3.8 27B Uncensored FP8	1.0M	$0.84/M	$2.63/M	$0.18/M	89.8%	3.5s	90.7 tps
Qwen 3.8 27B Uncensored Thinking FP8	1.0M	$0.84/M	$2.63/M	$0.18/M	79.5%	3.4s	86.7 tps
Qwen3 14B INT4	8k	$0.08/M	$0.23/M	$0.04/M	N/A	N/A	N/A
Qwen3 30B A3B Instruct 2507 INT4	49k	$0.19/M	$0.50/M	$0.09/M	N/A	N/A	N/A
Qwen3 Coder 30B A3B Instruct INT4	49k	$0.09/M	$0.40/M	$0.05/M	N/A	N/A	N/A
Qwen3.5 9B INT4	8k	$0.05/M	$0.14/M	$0.02/M	N/A	N/A	N/A
Qwen3.8 27B	262k	$0.22/M	$1.32/M	$0.04/M	N/A	N/A	N/A
Qwen3.8 27B Thinking	262k	$0.22/M	$1.32/M	$0.04/M	N/A	N/A	N/A
Schattenblume 31B INT4	66k	$0.11/M	$0.47/M	$0.05/M	N/A	N/A	N/A
The Drummer Cydonia 24B v4.3 INT4	66k	$0.11/M	$0.14/M	$0.06/M	N/A	N/A	N/A
TheDrummer Skyfall 31B v4.2 NVFP4	16k	$0.11/M	$0.14/M	$0.06/M	N/A	N/A	N/A

DeepSeek V4.1 Flash

Context: 262k
Input /M: $0.05/M
Output /M: $0.11/M
Cache read /M: $0.0008/M
Cache hit: N/A
TTFT: N/A
Throughput: N/A

DeepSeek V4.1 Flash (Thinking)

Context: 262k
Input /M: $0.05/M
Output /M: $0.11/M
Cache read /M: $0.0008/M
Cache hit: N/A
TTFT: N/A
Throughput: N/A

Gemma 3 12B IT

INT4

Context: 8k
Input /M: $0.26/M
Output /M: $0.26/M
Cache read /M: $0.13/M
Cache hit: N/A
TTFT: N/A
Throughput: N/A

Gemma 3 4B IT

INT4

Context: 8k
Input /M: $0.19/M
Output /M: $0.19/M
Cache read /M: $0.09/M
Cache hit: N/A
TTFT: N/A
Throughput: N/A

Gemma 4 12B Instruct

INT4

Context: 33k
Input /M: $0.05/M
Output /M: $0.24/M
Cache read /M: $0.02/M
Cache hit: N/A
TTFT: N/A
Throughput: N/A

Gemma 4 31B Gembrain uncensored heretic

INT4

Context: 66k
Input /M: $0.31/M
Output /M: $0.31/M
Cache read /M: $0.05/M
Cache hit: N/A
TTFT: N/A
Throughput: N/A

Gemma 4 31B Gembrain X Core

INT4

Context: 66k
Input /M: $0.31/M
Output /M: $0.31/M
Cache read /M: $0.05/M
Cache hit: N/A
TTFT: N/A
Throughput: N/A

Gemma 4 31B MeroMero v2

INT4

Context: 66k
Input /M: $0.09/M
Output /M: $0.46/M
Cache read /M: $0.05/M
Cache hit: 48.1%
TTFT: 2.5s
Throughput: 12.5 tps

Gemma 4 31B Split-Untied

Context: 262k
Input /M: $0.11/M
Output /M: $0.31/M
Cache read /M: N/A
Cache hit: 0%
TTFT: 1.3s
Throughput: 39.2 tps

Gemma 4 Dark Thoughts V2 31B

INT4

Context: 66k
Input /M: $0.11/M
Output /M: $0.47/M
Cache read /M: $0.05/M
Cache hit: N/A
TTFT: N/A
Throughput: N/A

Gemma 4 Giftige-Blume V2 31B

INT4

Context: 66k
Input /M: $0.11/M
Output /M: $0.47/M
Cache read /M: $0.05/M
Cache hit: N/A
TTFT: N/A
Throughput: N/A

Gemma 4 Gutenberg 31B

INT4

Context: 66k
Input /M: $0.11/M
Output /M: $0.47/M
Cache read /M: $0.05/M
Cache hit: N/A
TTFT: N/A
Throughput: N/A

Gemsicle

INT4

Context: 66k
Input /M: $0.09/M
Output /M: $0.46/M
Cache read /M: $0.05/M
Cache hit: N/A
TTFT: N/A
Throughput: N/A

GLM 5.3

Context: 1M
Input /M: $0.58/M
Output /M: $2.10/M
Cache read /M: $0.12/M
Cache hit: 49.5%
TTFT: 2.8s
Throughput: 148.5 tps

GLM 5.3 (Thinking)

Context: 1M
Input /M: $0.58/M
Output /M: $2.10/M
Cache read /M: $0.12/M
Cache hit: 49.6%
TTFT: 2.3s
Throughput: 111.6 tps

GLM 5.3 Flash Uncensored

NVFP4

Context: 524k
Input /M: $0.21/M
Output /M: $0.84/M
Cache read /M: $0.01/M
Cache hit: 3.9%
TTFT: 17.6s
Throughput: 24.4 tps

GrayLine Qwen3 8B

Context: 33k
Input /M: $0.29/M
Output /M: $0.29/M
Cache read /M: $0.15/M
Cache hit: N/A
TTFT: N/A
Throughput: N/A

LFM2.5 2.6B

Context: 8k
Input /M: $0.09/M
Output /M: $0.19/M
Cache read /M: $0.05/M
Cache hit: N/A
TTFT: N/A
Throughput: N/A

Llama 3.1 8B

INT4

Context: 8k
Input /M: $0.05/M
Output /M: $0.08/M
Cache read /M: $0.03/M
Cache hit: N/A
TTFT: N/A
Throughput: N/A

Llama 3.2 3B Instruct

INT4

Context: 8k
Input /M: $0.03/M
Output /M: $0.05/M
Cache read /M: $0.01/M
Cache hit: N/A
TTFT: N/A
Throughput: N/A

MiMo V2.6 Pro

Context: 1.0M
Input /M: $0.54/M
Output /M: $1.10/M
Cache read /M: $0.0045/M
Cache hit: 32.3%
TTFT: 2.8s
Throughput: 23.4 tps

Ministral 3B

INT4

Context: 8k
Input /M: $0.09/M
Output /M: $0.09/M
Cache read /M: $0.05/M
Cache hit: N/A
TTFT: N/A
Throughput: N/A

Ministral 8B

INT4

Context: 8k
Input /M: $0.14/M
Output /M: $0.14/M
Cache read /M: $0.07/M
Cache hit: N/A
TTFT: N/A
Throughput: N/A

Mistral Nemo

INT4

Context: 8k
Input /M: $0.09/M
Output /M: $0.11/M
Cache read /M: $0.05/M
Cache hit: N/A
TTFT: N/A
Throughput: N/A

Mistral Small 3 24B (2501)

INT4

Context: 8k
Input /M: $0.11/M
Output /M: $0.34/M
Cache read /M: $0.05/M
Cache hit: N/A
TTFT: N/A
Throughput: N/A

Mistral Small 3.2 24B (2506)

INT4

Context: 8k
Input /M: $0.19/M
Output /M: $0.40/M
Cache read /M: $0.09/M
Cache hit: N/A
TTFT: N/A
Throughput: N/A

Neural Daredevil 8B abliterated

INT4

Context: 8k
Input /M: $0.44/M
Output /M: $0.44/M
Cache read /M: $0.22/M
Cache hit: N/A
TTFT: N/A
Throughput: N/A

Phi 4 Mini

INT4

Context: 8k
Input /M: $0.16/M
Output /M: $0.69/M
Cache read /M: $0.08/M
Cache hit: N/A
TTFT: N/A
Throughput: N/A

Qwen 3.8 27B Uncensored

FP8

Context: 1.0M
Input /M: $0.84/M
Output /M: $2.63/M
Cache read /M: $0.18/M
Cache hit: 89.8%
TTFT: 3.5s
Throughput: 90.7 tps

Qwen 3.8 27B Uncensored Thinking

FP8

Context: 1.0M
Input /M: $0.84/M
Output /M: $2.63/M
Cache read /M: $0.18/M
Cache hit: 79.5%
TTFT: 3.4s
Throughput: 86.7 tps

Qwen3 14B

INT4

Context: 8k
Input /M: $0.08/M
Output /M: $0.23/M
Cache read /M: $0.04/M
Cache hit: N/A
TTFT: N/A
Throughput: N/A

Qwen3 30B A3B Instruct 2507

INT4

Context: 49k
Input /M: $0.19/M
Output /M: $0.50/M
Cache read /M: $0.09/M
Cache hit: N/A
TTFT: N/A
Throughput: N/A

Qwen3 Coder 30B A3B Instruct

INT4

Context: 49k
Input /M: $0.09/M
Output /M: $0.40/M
Cache read /M: $0.05/M
Cache hit: N/A
TTFT: N/A
Throughput: N/A

Qwen3.5 9B

INT4

Context: 8k
Input /M: $0.05/M
Output /M: $0.14/M
Cache read /M: $0.02/M
Cache hit: N/A
TTFT: N/A
Throughput: N/A

Qwen3.8 27B

Context: 262k
Input /M: $0.22/M
Output /M: $1.32/M
Cache read /M: $0.04/M
Cache hit: N/A
TTFT: N/A
Throughput: N/A

Qwen3.8 27B Thinking

Context: 262k
Input /M: $0.22/M
Output /M: $1.32/M
Cache read /M: $0.04/M
Cache hit: N/A
TTFT: N/A
Throughput: N/A

Schattenblume 31B

INT4

Context: 66k
Input /M: $0.11/M
Output /M: $0.47/M
Cache read /M: $0.05/M
Cache hit: N/A
TTFT: N/A
Throughput: N/A

The Drummer Cydonia 24B v4.3

INT4

Context: 66k
Input /M: $0.11/M
Output /M: $0.14/M
Cache read /M: $0.06/M
Cache hit: N/A
TTFT: N/A
Throughput: N/A

TheDrummer Skyfall 31B v4.2

NVFP4

Context: 16k
Input /M: $0.11/M
Output /M: $0.14/M
Cache read /M: $0.06/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 Kitani via the API

Call a model served by Kitani using its displayed model ID, including the provider suffix when selection is supported.

cURL

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:kitani",
    "messages": [{"role": "user", "content": "Hello"}]
  }'

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