Compare models

Put up to 4 models beside each other — token prices, context windows, capabilities and provider, from the same catalogue the model pages read.

Aura-2Deepgram Remove
Perceptron Mk1.5Perceptron Remove
Gemini 3.5 TranscribeGoogle Remove
Grok 4.7SpaceXAI Remove

4 is the maximum. Remove one to add another.

aura-2 vs perceptron-mk1.5 vs gemini-3.5-transcribe vs grok-4.7
Attribute	Aura-2aura-2	Perceptron Mk1.5perceptron-mk1.5	Gemini 3.5 Transcribegemini-3.5-transcribe	Grok 4.7grok-4.7
Pricing
Input	$0 / 1M	$0.15 / 1M	$0 / 1M	$1.60 / 1M
Output	$0 / 1M	$1.50 / 1M	$0 / 1M	$4.80 / 1M
Cache Write (5m)	Not applicable	$0.15 / 1M	Not applicable	$1.60 / 1M
Cache Write (1h)	Not applicable	$0.15 / 1M	Not applicable	$1.60 / 1M
Cache Read	Not applicable	$0.15 / 1M	Not applicable	$1.60 / 1M
Web Search	$0 / 1M	$0 / 1M	$0 / 1M	$0 / 1M
Context
Max context	N/A	36.9K	98.3K	500K
Max output	N/A	N/A	N/A	N/A
Capabilities
Vision	No	No	Yes	Yes
Function Calling	No	No	Yes	Yes
JSON Mode	No	No	No	Yes
Streaming	No	Yes	No	Yes
Catalogue
Provider	Deepgram	Perceptron	Google	SpaceXAI
Category	voice	chat	voice	chat
Charge type	Pay As You Go	Pay As You Go	Pay As You Go	Pay As You Go
Released	2026-07-16	2026-09-25	2026-09-25	—
Description
Summary	Deepgram Aura-2 text-to-speech for real-time voice applications. Billed per input character.	Perceptron Mk1.5 chat model that accepts audio input. Audio and text input are billed per token.	Google Gemini 3.5 Transcribe speech-to-text. Billed per input and output token.	Grok 4.7 is SpaceXAI's flagship model for coding, agentic workflows, and professional knowledge work, succeeding Grok 4.6. It is particularly strong at long-running software engineering, self-verification, and long-context execution, while improving capabilities in document drafting, presentations, and other professional tasks. Trained with extended reinforcement learning focused on multi-hour problems, Grok 4.7 is optimized for sustained, complex task execution and natively supports the Grok Bot harness for conversational workflows. It also introduces an enhanced safeguard stack designed to combine strong jailbreak resistance with low refusal rates for legitimate technical work. Reported benchmark results use xhigh reasoning effort.