Perceptron Mk1 API Pricing
Perceptron Mk1 is Perceptron's first flagship vision-language model, launched May 12, 2026: image and video in, text and grounded annotations out, with reasoning. Perceptron prices it at $0.15/M input, $1.50/M output and $0.0375/M cached input. Its successor Mk1.5 costs the same and adds audio and function calling. Checked by hand against docs.perceptron.inc.
Run the numbers.
Live calculator pre-loaded with Mk1's rates. Images and video frames bill as input tokens, so size the input side from Perceptron's tokenization guide; the cached rate applies to context reused across Multilook prompts.
Real-world presets.
Shelf image object boxes
Clip event timing
Receipt field extraction
Visual question
This is a chars-per-token approximation, not a real tokenizer. Actual tokens vary by language, code density, and tool-call overhead — counts are typically ±10–20% off for English prose, more for code or non-Latin scripts. For exact billing, use the vendor's official tokenizer.
| Model | Input /M | Output /M | Effective blended | Context | Best for |
|---|---|---|---|---|---|
| Perceptron Mk1 Current | $0.15 cache $0.0375 | $1.50 | $0.173 agentic 92/8 | 32K | Image and video perception with reasoning |
| Perceptron Mk1.5 | $0.15 cache $0.0375 | $1.50 | $0.173 same | 36K | Newer Mk, same price, adds audio |
| Gemini 3.5 Flash-Lite | $0.30 cache $0.03 | $2.50 | $0.272 +57% | 1M | Budget general multimodal model |
| Step 3.7 Flash | $0.20 cache $0.04 | $1.15 | $0.155 −10% | 256K | Video-capable reasoning model |
| Doubao Seed 2.1 Lite | $0.119 cache $0.0238 | $0.402 | $0.0698 −2.5× | 1M | Low-cost omni-modal model |
Frequently asked.
Perceptron Mk1 pricing questions - the rate, when the cache rate arrived, and whether to move to Mk1.5.
Q · 01 What is Perceptron Mk1 priced at? +
perceptron-mk1 at $0.15/M input, $1.50/M output and $0.0375/M cached input. Images and video frames bill as input tokens.Q · 02 Has the price changed since launch? +
$0.15/$1.50. The cached rate is newer - it came with the Multilook endpoint on September 3, 2026, which prefills a shared context once and bills reused tokens at $0.0375/M.Q · 03 Should I move to Mk1.5? +
reasoning_effort control, with a 36,864-token context against Mk1's 32K. Perceptron has not published a retirement date for Mk1; its MCP tools now default to Mk1.5 but still accept perceptron-mk1.Q · 04 How do I turn reasoning on or off? +
vision_config.enable_thinking: Perceptron recommends it on for text Q&A, captioning, OCR and video clipping, and off for spatial detection with points, boxes or polygons. Reasoning tokens are output tokens, so leaving it on for detection raises the bill for no gain.