Frontier Models Dashboard

Live metrics and curated analytics for the best models.

Top MultimodalGemini 3.1 ProGoogle

Best CodingGPT-5.4 ThinkingOpenAI

Top Open WeightsDeepSeek R2Deepseek

Fast Reasoningo3-miniOpenAI

Click to select and compare frontier models head-to-head.

Provider

Type

Model	Features		MMLU	HumanEval	Context (k)	TTFT (s)	Speed (Tk/s)	Input/1M	Output/1M
GPT-5.4 ThinkingOpenAISOTA Leader	PROP	1390	94.5%	96%	1050	0.9	65	$30.00	$180.00
Gemini 3.1 ProGoogleTop Multimodal	PROP	1385	93%	95.5%	1049	0.7	85	$2.00	$12.00
Claude 3.7 OpusAnthropicBest Reasoning	PROP	1380	92.5%	94.5%	1000	1	60	$5.00	$25.00
DeepSeek R2DeepseekTop Open Weights	OPEN	1365	90.5%	93.5%	131	0.8	70	$0.25	$0.38
Claude 3.7 SonnetAnthropicArena Leader	PROP	1360	89.4%	92%	200	0.8	75	$3.00	$15.00
o1OpenAITop Reasoning	PROP	1350	90.8%	92.4%	200	2.5	45	$15.00	$60.00
Grok 3xAISOTA	PROP	1345	88%	90.5%	131	0.6	85	$3.00	$15.00
DeepSeek R1DeepseekTop Open Weights	OPEN	1340	89%	91.5%	64	0.9	60	$0.70	$2.50
Gemini 2.5 ProGoogle	PROP	1335	88.5%	90%	1049	0.7	80	$1.25	$10.00
o3-miniOpenAIFast Reasoning	PROP	1330	87.5%	90.2%	200	0.4	120	$1.10	$4.40
DeepSeek V3Deepseek	OPEN	1300	87.5%	89%	164	0.5	110	$0.32	$0.89
GPT-4oOpenAIVersatile	PROP	1285	88.7%	90.2%	128	0.5	105	$2.50	$10.00
Qwen MaxQwen	PROP	1280	87%	88.5%	33	0.7	90	$1.04	$4.16
Llama 3.3 70BMetaTop Open Source	OPEN	1265	85.5%	81.5%	131	0.6	95	$0.12	$0.38