Models

Compare	Model	Context	Input /M	Output /M	Cached /M	Precision	Throughput	Latency	Capabilities	Uptime
	Kimi K3.1 Pro Moonshot	1M	$2.70	$13.50	$0.27		39tok/s	3.34s		100%
	MiniMax M3.1 Flash MiniMax	1M	$0.2025	$0.81	$0.02025		75tok/s	1.05s	+2	100%
	Muse Spark Open Meta	1M	$1.688	$5.738	—		160tok/s	1.33s		100%
	Nemotron 3.5 Super VL NVIDIA	1M	$0.081	$0.4725	$0.0135		123tok/s	0.90s	+1	100%
	Gemini 4 Argon50% off Google	1M	$2.70 $5.40	$13.50 $27.00	$0.135		38tok/sfrom traffic, 7d	8.50sfrom traffic, 7d	+3	96.3%
	Muse Wonder Meta	1M	$5.40	$27.00	—				+3
	Step 5 Preview StepFun	1M	$1.35	$3.645	$0.135		132tok/s	1.06s		100%
	Claude Haiku 5.5 Anthropic	1M	$0.135	$0.675	$0.0135		53tok/s	0.84s	+3	100%
	GPT-6.1 Sol OpenAI	1M	$2.70	$13.50	$0.135		36tok/s	1.34s	+2	100%
	Claude Sonnet 5.5 Anthropic	1M	$2.70	$13.50	$0.135		46tok/s	0.86s	+2	100%
	Claude Opus 5.5 Anthropic	1M	$5.40	$27.00	$0.27		27tok/s	1.32s	+2	100%
	GPT-6 Luna OpenAI	1M	$0.1639	$0.8195	$0.0135		45tok/s	1.40s	+3	100%
	Grok 4.7 xAI	500K	$2.16	$6.48	$0.54				+1
	MiMo V2.6 Pro Xiaomi	1M	$0.5805	$1.1745	—		69tok/s	0.75s		100%
	MiMo V2.6 Flash Xiaomi	1M	$0.189	$0.378	—		136tok/s	0.46s		100%
	Jev 1.13 TypeSafe	32K	$0.0567	$0.00	—
	DeepSeek V4.1 Flash51% off DeepSeek	1M	$0.1323 $0.27	$0.3969 $0.81	—		41tok/s	1.11s		100%
	GPT Image 2.5 Flare OpenAI	—	$0.019 / image				12s/img	11.60s
	Lyria 3.5 Google	—	$0.108 / song
	GPT-6 Astra OpenAI	1M	$13.50	$67.50	$1.35				+1
	Gemini 3.8 Flash50% off Google	1M	$1.013 $2.026	$5.063 $10.126	—		109tok/sfrom traffic, 7d	18.18sfrom traffic, 7d	+4	95.5%
	Muse Spark 1.3 Meta	1M	$1.688	$5.738	—				+3
	Claude Fable 5.1 Anthropic	1M	$13.50	$67.50	—
	Hy4 Preview Tencent	1M	$1.1259	$3.3764	—		43tok/s	8.70s		100%
	GPT Transcribe OpenAI	—	$0.006075 / min					1.55sfrom traffic, 7d		100%
	Gemini 3.5 Transcribe Google	—	$0.00675 / min							81.0%
	Qwen 3.8 Flash Alibaba	1M	$0.2045	$0.641	—		27tok/sfrom traffic, 7d	2.29sfrom traffic, 7d	+1	77.2%
	GLM 5.3 Flash Zhipu AI	1M	$0.172	$0.5733	—		61tok/sfrom traffic, 7d	3.01sfrom traffic, 7d	+1	99.9%
	DeepSeek V4 Flash Vision DeepSeek	1M	$0.3892	$1.168	—					98.4%
	GLM 5.3 Zhipu AI	1M	$1.178	$3.702	—
	Qwen 3.8 27B Alibaba	1M	$0.3826	$2.732	—					98.1%
	Gemini 3.7 Flash50% off Google	1M	$1.013 $2.026	$5.063 $10.126	$0.050625		32tok/sfrom traffic, 7d	4.47sfrom traffic, 7d	+5	84.3%
	Grok 4.6 xAI	500K	$2.758	$8.274	—
	Muse Glimmer 30B Meta	131K	$0.4554	$1.821	—					100%
	Muse Spark 1.2 Meta	1M	$1.773	$6.025	—				+3
	Qwen 3.8 Max Alibaba	1M	$2.228	$6.684	$0.2781
	MiMo V2.5 Xiaomi	1M	$0.1859	$0.3718	—
	Qwen 3.7 Flash Alibaba	1M	$0.0435	$0.1886	$0.0081		90tok/sfrom traffic, 7d	31.76sfrom traffic, 7d	+2	99.8%
	Claude Opus 5 Anthropic	1M	$6.75	$33.75	$0.675				+2
	Gemini 3.6 Flash Google	1M	$1.013	$5.063	$0.10125				+4	87.5%
	Gemini 3.5 Flash Lite Google	1M	$0.405	$3.375	$0.0405		48tok/s	1.29s	+4	99.6%
	Muse Spark 1.1 Meta	1M	$1.542	$5.243	—				+3
	Kimi K3 Moonshot	1M	$3.052	$15.258	$0.405		24tok/sfrom traffic, 7d	2.05sfrom traffic, 7d		100%
	GPT-5.6 Sol Pro50% off OpenAI	1M	$2.70 $5.40	$13.50 $27.00	$0.3375				+2
	GPT-5.6 Sol42% off OpenAI	1M	$3.151 $5.40	$15.753 $27.00	$0.393875		31tok/sfrom traffic, 7d	5.21sfrom traffic, 7d	+2	98.1%
	GPT-5.6 Terra Pro OpenAI	1M	$2.70	$16.20	$0.27				+2
	GPT-5.6 Terra OpenAI	1M	$2.70	$16.20	$0.27				+1
	GPT-5.6 Luna Pro OpenAI	1M	$0.2703	$1.621	$0.027				+3	100%
	GPT-5.6 Luna OpenAI	1M	$0.2944	$1.767	$0.027		69tok/sfrom traffic, 7d	7.54sfrom traffic, 7d	+3	95.3%
	Grok 4.5 xAI	500K	$2.73	$8.189	—				+1
	Hunyuan 3 Tencent	262K	$0.1594	$0.6576	—					100%
	Claude Sonnet 5 Anthropic	1M	$2.70	$13.50	$0.27		41tok/sfrom traffic, 7d	4.30sfrom traffic, 7d	+2	100%
	GLM 5.2 Zhipu AI	1M	$0.9029	$3.083	$0.29835	fp4 · fp8	62tok/s	4.08s		100%
	Kimi K2.7 Code Moonshot	262K	$1.009	$4.774	—
	Claude Fable 5 Anthropic	1M	$13.50	$67.50	—
	Nemotron 3 Ultra 550B NVIDIA	1M	$0.675	$3.375	$0.135	fp4 · fp8
	Qwen 3.7 Plus Alibaba	1M	$0.4325	$1.731	$0.0864					100%
	MiniMax M3 MiniMax	1M	$0.3956	$1.582	$0.0756	fp4 · fp8			+1	100%
	Step 3.7 Flash StepFun	256K	$0.2229	$1.282	$0.054	fp8			+2	100%
	Qwen 3.7 Max Alibaba	1M	$2.071	$6.207	$0.16875
	Grok Build xAI	256K	$1.389	$2.777	$0.27				+1
	Gemini 3.5 Flash Google	1M	$2.025	$12.15	$0.2025				+4	95.8%
	Grok 4.3 xAI	1M	$1.92	$3.838	$0.27
	DeepSeek V4 Flash DeepSeek	1M	$0.123	$0.2458	$0.0243		20tok/s	2.06s		99.9%
	GPT Image 2 OpenAI	—	$0.072 / image				36s/img	35.93s		97.7%
	Kimi K2.6 Moonshot	262K	$0.5909	$2.758	—		28tok/s	1.64s		100%
	Claude Opus 4.7 Anthropic	1M	$6.75	$33.75	$0.675				+1
	GLM 5.1 Zhipu AI	203K	$1.89	$5.94	$0.27675	fp4 · fp8
	Qwen 3.6 Plus Alibaba	1M	$0.4388	$2.633	—		51tok/s	9.78s		100%
	Gemma 4 31B Google	128K	$0.1642	$0.4689	$0.1642	fp4 · fp8	9tok/s	7.02s		99.7%
	Grok 4.20 Multi-Agent xAI	2M	$1.767	$3.534	—
	Grok 4.20 xAI	2M	$1.763	$3.528	—				+1
	MiniMax M2.7 MiniMax	205K	$0.405	$1.62	—					100%
	Nano Banana 3 Flash Google	—	$0.061 / image				10s/img	10.40s
	Gemini 3.1 Pro Google	1M	$2.70	$16.20	$0.27				+4
	Recraft V4 Vector Recraft	—	$0.108 / image
	Claude Sonnet 4.6 Anthropic	1M	$4.05	$20.25	$0.405				+1	100%
	Recraft V4 Recraft	—	$0.054 / image				21s/img	21.30s
	Recraft V4 Pro Recraft	—	$0.3375 / image
	MiniMax M2.5 MiniMax	197K	$0.4188	$1.675	—					98.4%
	Kling 3 Pro Kuaishou	—	$0.1512 / sec
	GLM 4.7 Flash Zhipu AI	203K	$0.081	$0.54	$0.0135	fp8				100%
	Gemini 3 Flash Google	1M	$0.675	$4.05	$0.0675		42tok/s	1.44s	+4	90.8%
	DeepSeek V3 DeepSeek	160K	$0.518	$0.7571	—	fp4
	Kling AI Avatar v2 Pro Kuaishou	—	$0.15525 / sec
	FLUX.2 Pro Black Forest Labs	—	$0.0405 / image
	Veo 3.1 Fast Google	—	$0.135 / sec
	Claude Haiku 4.5 Anthropic	200K	$1.35	$6.75	$0.135		172tok/sfrom traffic, 7d	4.85sfrom traffic, 7d	+2	100%
	EmbeddingGemma 300M Google	2K	$0.0027	$0.00	—
	GPT-OSS 120B OpenAI	128K	$0.0473	$0.2303	—					100%
	GLM 4.5 Air Zhipu AI	131K	$0.1797	$1.175	$0.03375	fp8				100%
	Qwen 3 Coder Alibaba	131K	$0.3546	$1.612	$0.135	fp4 · fp8	47tok/s	1.17s		96.2%
	Qwen3 Embedding 8B Alibaba	33K	$0.0135	$0.00	—					100%
	Qwen3 Embedding 0.6B Alibaba	33K	$0.0135	$0.00	—
	Qwen3 Embedding 4B Alibaba	33K	$0.027	$0.00	—
	Qwen3 Reranker 8B Alibaba	41K	$0.135	$0.00	—
	FLUX.1 Kontext Pro Black Forest Labs	—	$0.054 / image
	Qwen 3 32B Alibaba	33K	$0.108	$0.378	—					97.2%
	Llama 4 Maverick Meta	1M	$0.3344	$1.337	—	fp8				100%
	MiniMax Speech HD MiniMax	—	$0.07 / 1K char
	MiniMax Speech Turbo MiniMax	—	$0.04 / 1K char
	Ideogram V3 Ideogram	—	$0.081 / image				17s/img	16.68s
	Sonar Pro Perplexity	200K	$4.05	$20.25	—		4tok/sfrom traffic, 7d	6.28sfrom traffic, 7d
	Sonar Perplexity	127K	$1.35	$1.35	—		18tok/s	2.10s		100%
	MiniMax Image 01 MiniMax	—	$0.005 / image				16s/img	16.24s		100%
	DeepSeek R1 DeepSeek	64K	$0.7425	$2.957	—		55tok/s	1.60s		100%
	Llama 3.3 70B Meta	128K	$0.135	$0.432	—		13tok/s	1.10s		100%
	ElevenLabs Flash v2.5 ElevenLabs	—	$0.2025 / 1K char
	Whisper Large v3 Turbo OpenAI	—	$0.0009 / min					0.28sfrom traffic, 7d		99.3%
	ElevenLabs Turbo v2.5 ElevenLabs	—	$0.2025 / 1K char
	GPT-4o mini Transcribe OpenAI	—	$0.00405 / min					0.69sfrom traffic, 7d		99.6%
	Multilingual E5 Large Microsoft	1K	$0.0135	$0.00	—
	BGE-M3 BAAI	8K	$0.0135	$0.00	—
	BGE Large EN v1.5 BAAI	1K	$0.0135	$0.00	—
	BGE Base EN v1.5 BAAI	1K	$0.00675	$0.00	—
	ElevenLabs Multilingual v2 ElevenLabs	—	$0.405 / 1K char					1.54sfrom traffic, 7d		100%
	GTE Base Alibaba	1K	$0.00675	$0.00	—
	all-MiniLM-L12-v2 Sentence Transformers	1K	$0.00675	$0.00	—
	all-MiniLM-L6-v2 Sentence Transformers	1K	$0.00675	$0.00	—

Token models priced per 1M tokens. The cached column is what a repeated prompt prefix costs on that model — a rate of its own, read from the same source as the input rate beside it, not a fixed fraction of it. It runs from a tenth of the input rate to the full input rate depending on the model, and a dash means nobody publishes one, which is a different statement from “free”. Image, video, and audio models priced per unit (image, second, video, minute, song, call, or 1K characters). Throughput and latency are a probe with an identical input when Kyma has that sample. If a row is labeled “from traffic, 7d”, that number is production traffic instead — real prompts, not comparable across models. Image throughput is seconds per succeeded job (30-day median). Video, music and realtime stay empty until a job exists. An empty throughput, latency or uptime cell means Kyma does not yet have a number for that cell. Uptime is a percentage only after 20 observations, over 30 days, probes and real requests together, including Kyma's own accounts. Rankings demand is the same total. Precision names the numeric formats a model may be served at that are BELOW the one its creator released it in, and it is filled on the 14 models where that is true. A low format is not a downgrade when it is how the model shipped, so this is a comparison against the creator’s own model card rather than a list of small numbers. An empty precision cell means there is nothing to flag: either no route reports going under the release, or nobody could source what the release was. Those two are different answers and the model’s own page prints which one it is.