Tuna.ai

Models & Pricing

Model Input types Input /M Cache read /M Output /M SLA speed Parameter size Intelligence
Kimi K3 Ultrafast Image, Video $18.00 $1.80 $90.00 >250 tok/s 2.8 trillion 4 out of 4
DeepSeek V4.1 Flash Ultrafast Image $1.80 $0.18 $7.20 >350 tok/s 552 billion 2 out of 4
GLM-5.3 Flash Ultrafast Image, Video $0.90 $0.18 $3.00 >450 tok/s 320 billion 2 out of 4
Automatically receive up to 50% off if the output speed does not meet SLA.

Tuna currently only serves open-weight models, with US regional routing and Zero Data Retention.

Prices shown are denominated in US dollars, per 1 million tokens. Precision of billing is per 1 single token.

Use the specific model ID to pin your requests to that exact model. If that model is deprecated, requests to its ID will fail. The auto-updating model ID routes to the latest model in the same family, so you can use the newest model without changing your code. Prices can change between model generations in a family. When using an auto-updating model ID, billing is based on whichever model you are routed to.

Cache read is the price for cached input tokens. Input is implicitly a cache write, and Tuna bills input and cache writes at the same rate. See Billing for how token usage is charged.

Kimi K3 Ultrafast

Our strongest open-weight model, from Moonshot AI.

Max output
131,072 tokens

The output may be shorter to keep input and output within the model's total context length. Setting max_tokens above this limit won't increase the maximum output.

Reasoning effort levels
low high max
Active parameters
104 billion
Image pixels per token
28x28 pixels = 1 token
  • US Regional Routing
  • Precision: Original weights
  • 2.8 trillion parameters
  • SLA speed: >250 tok/s
Auto-updating model ID
tuna-us/moonshotai/kimi-latest-ultrafast
Specific model ID
tuna-us/moonshotai/kimi-k3-ultrafast
Input Cache read Output
$18.00 $1.80 $90.00

DeepSeek V4.1 Flash Ultrafast

A fast and economical workhorse from DeepSeek. According to DeepSeek, V4.1 Flash is more powerful than the much-larger V4 Pro.

Max output
131,072 tokens

The output may be shorter to keep input and output within the model's total context length. Setting max_tokens above this limit won't increase the maximum output.

Reasoning effort levels
none low high max
Active parameters
8 billion for input, 16 billion for output
Image pixels per token
42x42 pixels = 1 token
Auto-updating model ID
tuna-us/deepseek/deepseek-flash-latest-ultrafast
Specific model ID
tuna-us/deepseek/deepseek-v4.1-flash-ultrafast
Input Cache read Output
$1.80 $0.18 $7.20

GLM-5.3 Flash Ultrafast

A fast, low-cost model from Z.ai with moderate intelligence.

Max output
131,072 tokens

The output may be shorter to keep input and output within the model's total context length. Setting max_tokens above this limit won't increase the maximum output.

Reasoning effort levels
low high max
Active parameters
18 billion
Image pixels per token
28x28 pixels = 1 token
Auto-updating model ID
tuna-us/z-ai/glm-flash-latest-ultrafast
Specific model ID
tuna-us/z-ai/glm-5.3-flash-ultrafast
Input Cache read Output
$0.90 $0.18 $3.00