Does quality drop near the 1M limit?

Yes, usually. Most 1M-context models lose accuracy on needle-in-haystack tests past 500K tokens. Gemini 2.5 Pro holds up best on long-context benchmarks. For critical work past 500K, test before relying on it.

Is 1M context cheaper than chunked RAG?

Not always. Sending a 1M-token prompt can cost $1 to $15 per call. A well-tuned RAG with 8K context and good retrieval often wins on both cost and accuracy. Use 1M context when retrieval quality is the bottleneck.

How much does a single 1M-token call cost?

Roughly $-1000000.00 on the cheap end and $0.40 on the premium end per 1M input tokens. Output bills on top at the model's output rate.

Can I cache a 1M-token context?

Yes. Anthropic, Google, and OpenAI all support context caching. This is the single biggest cost lever for long-context work: cached reads typically cost 10 to 30 percent of a fresh read.

Context · 1M+

Cheapest 1M context LLMs

Every LLM with a 1,000,000+ token context window. Ranked by input price per 1M tokens.

Models40

Cheapest$-1000000.00

Min context1M tokens

All pricing Pricing home

What this page is

This page lists every priced model with a context window of at least one million tokens. 1M context unlocks whole-repo coding, book-length analysis, and massive multi-document RAG without chunking. The cost per call can be steep, so compare carefully and lean on context caching whenever possible.

Ranked by input price

1M+ context models, cheapest first.

#	Model	Provider	In $/1M	Out $/1M	Context	context	Type
1	Auto Router	openrouter	$-1000000.00	$-1000000.00	2.0M	2.0M	Closed
2	Fusion	openrouter	$-1000000.00	$-1000000.00	1.0M	1.0M	Closed
3	Pareto Code Router	openrouter	$-1000000.00	$-1000000.00	2.0M	2.0M	Closed
4	Lyria 3 Clip Preview	Google DeepMind	$0.00	$0.00	1.0M	1.0M	Closed
5	Lyria 3 Pro Preview	Google DeepMind	$0.00	$0.00	1.0M	1.0M	Closed
6	Nemotron 3 Super (free)	NVIDIA	$0.00	$0.00	1.0M	1.0M	OSS
7	Nemotron 3 Ultra (free)	NVIDIA	$0.00	$0.00	1.0M	1.0M	OSS
8	Owl Alpha	openrouter	$0.00	$0.00	1.0M	1.0M	Closed
9	Qwen3 Coder 480B A35B (free)	Alibaba Qwen	$0.00	$0.00	1.0M	1.0M	OSS
10	Qwen3.6 Plus (free)	Alibaba Qwen	$0.00	$0.00	1.0M	1.0M	Closed
11	Qwen3.6 Plus Preview (free)	Alibaba Qwen	$0.00	$0.00	1.0M	1.0M	OSS
12	Qwen3.5-Flash	Alibaba Qwen	$0.07	$0.26	1.0M	1.0M	OSS
13	Gemini 2.0 Flash Lite	Google DeepMind	$0.07	$0.30	1.0M	1.0M	Closed
14	Nemotron 3 Super	NVIDIA	$0.08	$0.45	1.0M	1.0M	OSS
15	DeepSeek V4 Flash	DeepSeek	$0.08	$0.17	1.0M	1.0M	OSS
16	Gemini 2.0 Flash	Google DeepMind	$0.10	$0.40	1.0M	1.0M	Closed
17	Gemini 2.5 Flash Lite	Google DeepMind	$0.10	$0.40	1.0M	1.0M	Closed
18	Gemini 2.5 Flash Lite Preview 09-2025	Google DeepMind	$0.10	$0.40	1.0M	1.0M	Closed
19	GPT-4.1 Nano	OpenAI	$0.10	$0.40	1.0M	1.0M	Closed
20	Llama 4 Scout	Meta	$0.10	$0.30	10.0M	10.0M	OSS
21	MiMo-V2.5	xiaomi	$0.10	$0.28	1.0M	1.0M	OSS
22	Llama 4 Maverick	Meta	$0.15	$0.60	1.0M	1.0M	OSS
23	Qwen3.6 Flash	Alibaba Qwen	$0.19	$1.13	1.0M	1.0M	OSS
24	Qwen3 Coder Flash	Alibaba Qwen	$0.20	$0.97	1.0M	1.0M	OSS
25	Grok 4.1 Fast	xAI	$0.20	$0.50	2.0M	2.0M	Closed
26	MiniMax-01	minimax	$0.20	$1.10	1.0M	1.0M	OSS
27	Qwen3 Coder 480B A35B	Alibaba Qwen	$0.22	$1.80	1.0M	1.0M	OSS
28	Gemini 3.1 Flash Lite	Google DeepMind	$0.25	$1.50	1.0M	1.0M	Closed
29	Gemini 3.1 Flash Lite Preview	Google DeepMind	$0.25	$1.50	1.0M	1.0M	Closed
30	Qwen Plus 0728	Alibaba Qwen	$0.26	$0.78	1.0M	1.0M	OSS
31	Qwen Plus 0728 (thinking)	Alibaba Qwen	$0.26	$0.78	1.0M	1.0M	OSS
32	Qwen-Plus	Alibaba Qwen	$0.26	$0.78	1.0M	1.0M	OSS
33	Qwen3.5 Plus 2026-02-15	Alibaba Qwen	$0.26	$1.56	1.0M	1.0M	OSS
34	Gemini 2.5 Flash	Google DeepMind	$0.30	$2.50	1.0M	1.0M	Closed
35	MiniMax M3	minimax	$0.30	$1.20	1.0M	1.0M	OSS
36	Nova 2 Lite	Amazon	$0.30	$2.50	1.0M	1.0M	Closed
37	Qwen3.5 Plus 2026-04-20	Alibaba Qwen	$0.30	$1.80	1.0M	1.0M	OSS
38	Qwen3.7 Plus	Alibaba Qwen	$0.32	$1.28	1.0M	1.0M	OSS
39	Qwen3.6 Plus	Alibaba Qwen	$0.33	$1.95	1.0M	1.0M	OSS
40	GPT-4.1 Mini	OpenAI	$0.40	$1.60	1.0M	1.0M	Closed

Top 3 cheapest 1M context LLMs

Auto Router delivers a 2.0M context window at $-1000000.00/M input tokens. Suitable for whole-codebase analysis, long legal docs, and book-scale RAG.

Fusion delivers a 1.0M context window at $-1000000.00/M input tokens. Suitable for whole-codebase analysis, long legal docs, and book-scale RAG.

Pareto Code Router delivers a 2.0M context window at $-1000000.00/M input tokens. Suitable for whole-codebase analysis, long legal docs, and book-scale RAG.

The price gap · cheapest vs most expensive

Cheapest

Auto Router

$-1000000.00/M

$ per 1M input tokens

Why the gap

Premium 1M-context models pay for better accuracy at the tail of the window and faster ingestion. For research and one-shot analysis, the cheap end delivers equivalent answers on most prompts.

Most expensive

GPT-4.1 Mini

$0.40/M

$ per 1M input tokens

Frequently asked questions

Gemini 2.5 Pro and Flash were first to a real 1M window. Claude Sonnet extended to 1M. Qwen3 Long is a strong open-source option. MiniMax and several Chinese labs also ship 1M+. See the table above for current live list.

Cheapest 1M context LLMs

Ranked by input price

Top 3 cheapest 1M context LLMs

The price gap · cheapest vs most expensive

Frequently asked questions

See also

Other context tiers

Stacks

Compare