Long-context AI models: 82 that hold a million tokens or more

Long context 1M+

Services found: 82

#	AI model	API $/1M (in / out)	Context window	Price / mo	Output $/1M	Overview
21	Qwen3.8 Max (0902) Qwen	$2 / $6	1M tokens	API only	6	View
22	Qwen3.8 2.4T A95B Qwen	$2 / $6	1.05M tokens	API only	6	View
23	Muse Spark 1.2 Meta	$1.25 / $4.25	1.05M tokens	API only	4.25	View
24	Gemini 3.7 Flash Google	$0.75 / $3.75	1.05M tokens	API only	3.75	View
25	GPT-5.5 OpenAI	$5 / $30	1.05M tokens	API only	30	View
26	Claude Sonnet 5 Anthropic	$2 / $10	1M tokens	API only	10	View
27	GPT-5.6 Luna OpenAI	$0.2 / $1.2	1.05M tokens	API only	1.2	View
28	DeepSeek V4 Pro 0813 DeepSeek	$1.0494 / $3.1482	1.05M tokens	API only	3.148	View
29	DeepSeek V4 Flash 0731 DeepSeek	$0.065 / $0.18	1.31M tokens	API only	0.18	View
30	Gemini 3.6 Flash Google	$0.75 / $3.75	1.05M tokens	API only	3.75	View
31	Muse Spark 1.1 Meta	$1.25 / $4.25	1.05M tokens	API only	4.25	View
32	Qwen3.8 27B Qwen	$0.42 / $3	1M tokens	API only	3	View
33	DeepSeek V4 Pro 0423 DeepSeek	$0.95526 / $1.91052	1.05M tokens	API only	1.911	View
34	Claude Sonnet 4.6 Anthropic	$3 / $15	1M tokens	API only	15	View
35	Qwen3.7 Max Qwen	$2.5 / $7.5	1M tokens	API only	7.5	View
36	MiniMax M3 MiniMax	$0.3 / $1.2	1.05M tokens	API only	1.2	View
37	MiMo-V2.5-Pro Xiaomi	$0.435 / $0.87	1.05M tokens	API only	0.87	View
38	Inkling Small Thinkingmachines	$0.45 / $1.2	1.05M tokens	API only	1.2	View
39	Qwen3.7 Plus Qwen	$0.32 / $1.28	1M tokens	API only	1.28	View
40	Inkling Thinkingmachines	$1.87 / $4.68²	1.05M tokens	API only	4.68	View

Browse by need

Reasoning models169 models

Coding assistants17 models

Vision and images156 models

Long context 1M+82 models

Free tier available80 models

Open weights88 models

Eighty-eight models hold a million tokens or more in a single conversation - roughly 750,000 words, or an entire codebase, a full contract bundle, a multi-hour transcript. Qwen-Long stretches furthest at 10M tokens.

Long context is not free in either sense. Several vendors charge more above a threshold - Gemini 3.1 Pro doubles from $2 to $4 on input past 200K tokens. Attention quality also degrades across very long inputs, so retrieval over a well-built index often beats pasting everything in. Lean on prompt caching when the same document is queried repeatedly.

Bigger is not smarterAttention thins out across very long inputs on every model available today.

Thresholds cost moneySeveral vendors raise the rate above 200K tokens, doubling input price mid-request.

Retrieval still winsA good index usually beats pasting the whole corpus, at a fraction of the cost.

Models with this capability

The 82 models below all support this capability. Prices are input and output per 1M tokens from the vendor official price list; for dirhams multiply by 3.6725. Tap a card to open the model page.

Claude Fable 5.1

Anthropic

in $10out $501M tokens

GPT-6 Astra

OpenAI

in $10out $501.05M tokens

GPT-6.1 Sol

OpenAI

in $2out $101.05M tokens

GPT-6 Sol

OpenAI

in $2out $101.05M tokens

GPT-6 Luna

OpenAI

in $0.1out $0.51.05M tokens

Claude Opus 5.5

Anthropic

in $4out $201M tokens

Claude Sonnet 5.5

Anthropic

in $2out $101M tokens

DeepSeek V4.1 Flash

DeepSeek

in $0.3out $1.21M tokens

How to choose

Use the full window when the material must be seen at once — not as a substitute for retrieval.
Watch for price tiers above 200K and budget accordingly.
Combine with caching when the same long document is queried repeatedly.
Check the maximum output separately; a big input window does not imply a big answer.

Common mistakes

Pasting everything by defaultMost questions need a fraction of the context and answer better without the noise.

Ignoring the output ceilingLong analyses hit max output limits well before the input window fills.

Who this suits

Legal and complianceContract bundles and regulatory texts read in one pass.

EngineeringRepository-scale refactors where cross-file context matters.

MediaMulti-hour transcripts summarised without chunking.

Related filters

Prompt caching147 models with cheap repeated context

Vision156 models that read images

Reasoning169 models that think before answering

Frontier tier41 flagship models, the best each vendor sells

All 239 modelsThe full rating with filters and prices

Frequently asked questions

Is a bigger window always better?

No. Attention quality degrades across very long inputs, and retrieval over a good index often beats pasting everything in.

Does long context cost extra?

Often. Gemini 3.1 Pro doubles input from $2 to $4 above 200K tokens, and several vendors have similar thresholds.

What is the largest window available?

Qwen-Long at 10M tokens. Most current flagships sit at 1M.

What is the largest window on the market?

Qwen-Long at 10M tokens. Most flagships sit at 1M.

Does long context slow responses?

Yes, materially. Time to first token grows with input size.

Frequently searched

Ratings 0

Avg. score 0

Rate this page

Help us improve