Best AI models for long documents.
Which models actually use a long context window: reasoning across large document sets, plus blind votes on questions about uploaded files. As of 2 Oct 2026, Kimi K3 leads for long context, ahead of Step 5 and MiMo V2.6 Pro, out of 54 ranked models. Kimi K3 has open weights, so it can also be self-hosted. The cheapest model in the top ten is DeepSeek V4.1 Flash, at A$0.76 per million tokens.
| MODEL | IN AUSTRALIA | |||
|---|---|---|---|---|
| 01Kimi K3MOONSHOT AIPARTIAL | 100.0 | 88.7% | A$8.63 | GLOBAL |
| 02Step 5STEPFUNPARTIAL | 97.9 | 88.3% | A$2.05 | |
| 03MiMo V2.6 ProXIAOMIPARTIAL | 87.6 | 86.3% | A$0.78 | NOT IN AU |
| 04Claude Fable 5.1ANTHROPIC | 85.0 | 85.3% | A$28.78 | GLOBAL |
| 05Claude Opus 5.5ANTHROPICPARTIAL | 79.4 | 84.7% | A$11.51 | IN AU |
| 06GPT-5.6 SolOPENAI | 78.4 | 84.0% | A$11.51 | GLOBAL |
| 07Claude Fable 5ANTHROPIC | 76.9 | 82.3% | A$28.78 | GLOBAL |
| 08GPT-5.5OPENAI | 76.8 | 84.3% | A$16.19 | GLOBAL |
| 09DeepSeek V4.1 FlashDEEPSEEKPARTIAL | 75.8 | 84.0% | A$0.76 | SELF-HOST |
| 10GPT-6 SolOPENAIPARTIAL | 74.2 | 83.7% | A$5.76 | GLOBAL |
| 11GPT-6 LunaOPENAIPARTIAL | 72.2 | 83.3% | A$0.29 | GLOBAL |
| 12GPT-6.1 SolOPENAIPARTIAL | 70.6 | 83.0% | A$5.76 | GLOBAL |
| 13GPT-5.6 TerraOPENAI | 69.3 | 83.0% | A$6.48 | GLOBAL |
| 14Claude Sonnet 5.5ANTHROPICPARTIAL | 69.1 | 82.7% | A$5.76 | GLOBAL |
| 15Qwen3.8 27BALIBABAPARTIAL | 65.5 | 82.0% | A$1.62 | SELF-HOST |
| 16Claude Sonnet 5ANTHROPIC | 65.3 | 82.0% | A$5.76 | IN AU |
| 17Muse Spark 1.3META | 65.3 | 83.0% | A$2.88 | NOT IN AU |
| 18GPT-5.6 LunaOPENAI | 65.2 | 83.7% | A$0.65 | GLOBAL |
| 19Gemini 3.7 FlashGOOGLEPARTIAL | 63.9 | 81.7% | A$2.16 | GLOBAL |
| 20GPT-5.4OPENAI | 62.7 | 82.0% | A$8.09 | IN AU |
| 21Claude Opus 5ANTHROPIC | 61.9 | 79.3% | A$14.39 | IN AU |
| 22Gemini 3.8 FlashGOOGLEPARTIAL | 61.9 | 81.3% | A$2.16 | GLOBAL |
| 23Claude Opus 4.7ANTHROPIC | 61.8 | 78.7% | A$14.39 | IN AU |
| 24Claude Opus 4.6ANTHROPIC | 59.8 | 78.0% | A$14.39 | IN AU |
| 25GPT-6 AstraOPENAI | 59.2 | 80.7% | A$28.78 | GLOBAL |
| 26Claude Sonnet 4.6ANTHROPIC | 59.0 | 80.0% | A$8.63 | IN AU |
| 27Gemini 3.1 ProGOOGLE | 57.6 | 82.0% | A$6.48 | GLOBAL |
| 28DeepSeek V4 ProDEEPSEEKPARTIAL | 56.7 | 80.3% | A$2.85 | GLOBAL |
| 29Qwen3.8 MaxALIBABAPARTIAL | 56.7 | 80.3% | A$4.32 | NOT IN AU |
| 30GLM-5.3 FlashZ.AIPARTIAL | 55.2 | 80.0% | A$0.34 | SELF-HOST |
| 31Claude Opus 4.8ANTHROPIC | 54.4 | 77.7% | A$14.39 | IN AU |
| 32DeepSeek V4 FlashDEEPSEEKPARTIAL | 53.6 | 79.7% | A$0.95 | GLOBAL |
| 33Gemini 4 ArgonGOOGLEPARTIAL | 53.6 | 79.7% | A$5.76 | NOT IN AU |
| 34GLM-5.3Z.AIPARTIAL | 53.6 | 79.7% | A$3.09 | SELF-HOST |
| 35MiMo V2.5 ProXIAOMIPARTIAL | 53.6 | 79.7% | A$0.78 | SELF-HOST |
| 36Qwen3.8 Flash NextALIBABAPARTIAL | 53.6 | 79.7% | A$0.33 | SELF-HOST |
| 37Grok 4.6XAI | 52.5 | 80.3% | A$4.32 | GLOBAL |
| 38Muse Spark 1.1META | 51.5 | 77.7% | A$2.88 | |
| 39Grok 4.5XAI | 50.3 | 79.3% | A$4.32 | NOT IN AU |
| 40Hy3TENCENTPARTIAL | 50.0 | 79.0% | A$0.35 | SELF-HOST |
| 41Muse Spark 1.2METAPARTIAL | 50.0 | 79.0% | A$2.88 | NOT IN AU |
| 42Qwen3.7 MaxALIBABAPARTIAL | 50.0 | 79.0% | A$5.40 | |
| 43Kimi K2.6MOONSHOT AI | 49.9 | 81.0% | A$2.46 | SELF-HOST |
| 44MiniMax M3MINIMAX | 49.4 | 83.0% | A$0.76 | SELF-HOST |
| 45Gemini 3.6 FlashGOOGLE | 47.9 | 80.0% | A$2.16 | GLOBAL |
| 46Muse SparkMETA | 47.0 | 78.0% | ||
| 47GLM-5.2Z.AIPARTIAL | 46.4 | 78.3% | A$3.09 | SELF-HOST |
| 48GPT-5.4 miniOPENAIPARTIAL | 39.7 | 77.0% | A$2.43 | |
| 49Grok 4.7XAIPARTIAL | 38.1 | 76.7% | A$4.32 | NOT IN AU |
| 50Gemini 3.5 Flash-LiteGOOGLEPARTIAL | 34.5 | 76.0% | A$1.22 | GLOBAL |
| 51Gemini 3 ProGOOGLE | 32.4 | 76.0% | A$6.48 | |
| 52Gemini 3.5 FlashGOOGLE | 27.0 | 73.3% | A$4.86 | IN AU |
| 53Gemma 4 31BGOOGLE | 6.7 | 69.7% | A$0.29 | SELF-HOST |
| 54Mistral Medium 3.5MISTRAL AIPARTIAL | 0.0 | 69.3% | A$4.32 | SELF-HOST |
What goes into the long context score.
A big context window only helps if the model can find and connect what matters inside it. AA-LCR tests exactly that across roughly 100,000-token document sets; the document arena adds how people rate answers about their own files. Models with a documented window under 128K tokens are left out.
Artificial Analysis’s long-context reasoning test: questions that need information pulled together from sets of documents of around 100,000 tokens.
Blind votes on answers to questions about uploaded documents (style-controlled).
Scores are rescaled 0–100 across the models we track, so they compare models with each other rather than against a fixed bar. Full methodology →
Choosing a model for long context.
What is the best AI model for long context right now?
As of 2 Oct 2026, Kimi K3 ranks first, followed by Step 5 and MiMo V2.6 Pro. The ranking combines AA-LCR and LMArena Document and is refreshed weekly.
What is the best open-weights model for long context?
Kimi K3 from Moonshot AI is the highest-ranked open-weights model (1st of 54). Open weights can be run on your own infrastructure, including in an Australian region.
Which top long context model is cheapest?
Of the top ten, DeepSeek V4.1 Flash is cheapest at A$0.76 per million tokens (blended), against A$28.78 for Claude Fable 5.1. Prices are list prices converted at A$1 = US$0.6949, excluding GST.
Which of these models can keep data in Australia?
Of the top ten, Claude Opus 5.5 can run with requests kept in Australia on at least one major cloud platform. See the Australia page for each platform.
How is this ranking calculated?
A big context window only helps if the model can find and connect what matters inside it. AA-LCR tests exactly that across roughly 100,000-token document sets; the document arena adds how people rate answers about their own files. Models with a documented window under 128K tokens are left out. Each input is rescaled 0–100 across the models we track, then weighted: AA-LCR 70% and LMArena Document 30%.
Ratings: LMArena leaderboard dataset (CC BY 4.0), rescaled for the lens scores · leaderboards to 30 Sep 2026
Evaluations, prices and speed: Artificial Analysis (artificialanalysis.ai) · fetched 1 Oct 2026
Exchange rate: Reserve Bank of Australia, table F11.1 · A$1 = US$0.6949 on 1 Oct 2026 · prices exclude GST
Snapshot 2 Oct 2026 · updated weekly · How the rankings work →