Compare models

Context, limits, capabilities, price and use cases side by side — up to 4 models at a time. The URL holds your selection, so you can share it.

Grok 4.3Grok 4.5 Add a modelChat with these 2

AttributeHide identical rows	Grok 4.3x-ai/grok-4.3	Grok 4.5x-ai/grok-4.5
Specs
Context window	1M	500K
Max output	1M	500K
Input modalities	TextImage	TextImage
Output modalities	Text	Text
Capabilities
Streaming	Supported	Supported
Tool calling	Supported	Supported
Structured output (JSON)	Supported	Supported
Image input	Supported	Supported
Thinking mode	Supported	Supported
Price / 1M tokens
Input	$ 1.25>= 200K$ 2.50	$ 2.00>= 200K$ 4.00
Output	$ 2.50>= 200K$ 5.00	$ 6.00>= 200K$ 12.00
Cache read	$ 0.20>= 200K$ 0.40	$ 0.30>= 200K$ 0.60
Cache write	$ 1.25>= 200K$ 2.50	$ 2.00>= 200K$ 4.00
What it is good at
Best for	Large-context workflows Tool-assisted knowledge work	Agentic engineering Technical knowledge work
Published benchmarks
SWE Marathonpass_at_1	—	29%(setup not fully disclosed)
Terminal-Bench 2.1pass_ratesetups differ by: reasoning	39.7%effort high(setup not fully disclosed)	83.3%(setup not fully disclosed)
DeepSWE 1.0resolved_rate	—	62%provider(setup not fully disclosed)
DeepSWE 1.1resolved_rate	—	53%mini-swe-agent
SWE-Bench Proresolved_rate	—	64.7%(setup not fully disclosed)
SciCodescore	47.3%effort high(setup not fully disclosed)	—
GDPval-AA v2elo	1087.5effort high(setup not fully disclosed)	—
AA-LCRscore	66%effort high(setup not fully disclosed)	—
AA Intelligence Index v4.1.1score	37.9effort high(setup not fully disclosed)	—
Humanity's Last Examaccuracy	37.2%effort high(setup not fully disclosed)	—
CritPtscore	8%effort high(setup not fully disclosed)	—
GPQA Diamondaccuracy	90.1%effort high(setup not fully disclosed)	—
tau3-Bankingscore	12.4%effort high(setup not fully disclosed)	—