Best AI models for reasoning and analysis.

The models that handle hard, multi-step problems best, on independent tests rather than chat preferences. As of 2 Oct 2026, Claude Opus 5.5 leads for reasoning, ahead of Claude Sonnet 5.5 and Claude Fable 5.1, out of 54 ranked models. The highest-ranked open-weights model is Kimi K3 (14th). The cheapest model in the top ten is Muse Spark 1.3, at A$2.88 per million tokens.

54 MODELS RANKEDA$1 = US$0.6949UPDATED 2 OCT 2026
REASONING RANKING54 MODELS · 2 OCT 2026
54 MODELS
Best AI models for reasoning and analysis, ranked, as of 2 Oct 2026
MODELIN AUSTRALIA
01Claude Opus 5.5ANTHROPICPARTIAL100.057.661.4%A$11.51IN AU
02Claude Sonnet 5.5ANTHROPICPARTIAL93.156.055.0%A$5.76GLOBAL
03Claude Fable 5.1ANTHROPIC91.153.459.1%93.7%A$28.78GLOBAL
04GPT-6 AstraOPENAI90.852.754.7%96.1%A$28.78GLOBAL
05Gemini 4 ArgonGOOGLEPARTIAL89.352.657.1%A$5.76NOT IN AU
06Claude Opus 5ANTHROPIC85.350.854.9%93.2%A$14.39IN AU
07GPT-6.1 SolOPENAIPARTIAL85.151.852.9%A$5.76GLOBAL
08Claude Fable 5ANTHROPIC83.649.655.5%92.6%A$28.78GLOBAL
09Muse Spark 1.3META79.348.148.7%93.5%A$2.88NOT IN AU
10GPT-5.6 SolOPENAI79.247.049.5%94.1%A$11.51GLOBAL
11GPT-6 SolOPENAIPARTIAL75.247.647.9%A$5.76GLOBAL
12MiMo V2.6 ProXIAOMIPARTIAL74.246.349.4%A$0.78NOT IN AU
13Grok 4.6XAI73.644.342.9%94.9%A$4.32GLOBAL
14Kimi K3MOONSHOT AI73.243.646.9%93.5%A$8.63GLOBAL
15Gemini 3.8 FlashGOOGLE72.740.947.8%95.3%A$2.16GLOBAL
16Qwen3.8 MaxALIBABA72.545.443.1%92.8%A$4.32NOT IN AU
17Claude Opus 4.8ANTHROPIC70.341.848.7%92.0%A$14.39IN AU
18GLM-5.3Z.AI70.144.842.3%91.7%A$3.09SELF-HOST
19Grok 4.7XAIPARTIAL70.046.443.1%A$4.32NOT IN AU
20Gemini 3.7 FlashGOOGLE69.739.147.9%94.5%A$2.16GLOBAL
21GPT-5.6 TerraOPENAI68.242.142.9%92.5%A$6.48GLOBAL
22Step 5STEPFUNPARTIAL68.243.746.5%A$2.05
23GPT-5.5OPENAI66.638.445.8%93.5%A$16.19GLOBAL
24Claude Opus 4.7ANTHROPIC65.040.742.3%91.4%A$14.39IN AU
25Grok 4.5XAI65.038.842.7%93.1%A$4.32NOT IN AU
26GLM-5.3 FlashZ.AI64.841.839.9%91.2%A$0.34SELF-HOST
27GPT-5.4OPENAI64.539.043.7%92.0%A$8.09IN AU
28Muse Spark 1.2META64.239.645.5%90.4%A$2.88NOT IN AU
29Qwen3.8 Flash NextALIBABA62.739.838.0%92.3%A$0.33SELF-HOST
30Claude Sonnet 5ANTHROPIC61.238.241.3%91.1%A$5.76IN AU
31DeepSeek V4 ProDEEPSEEK60.536.041.0%92.8%A$2.85GLOBAL
32GPT-5.6 LunaOPENAI59.237.339.5%91.1%A$0.65GLOBAL
33Gemini 3.6 FlashGOOGLE58.134.040.8%92.8%A$2.16GLOBAL
34Gemini 3.1 ProGOOGLE57.929.747.0%94.1%A$6.48GLOBAL
35Muse Spark 1.1META57.133.746.2%89.8%A$2.88
36Gemini 3.5 FlashGOOGLE56.832.642.7%92.2%A$4.86IN AU
37DeepSeek V4.1 FlashDEEPSEEKPARTIAL56.739.539.2%A$0.76SELF-HOST
38DeepSeek V4 FlashDEEPSEEK55.034.338.6%90.8%A$0.95GLOBAL
39GLM-5.2Z.AI54.133.741.1%89.5%A$3.09SELF-HOST
40GPT-6 LunaOPENAIPARTIAL54.038.138.5%A$0.29GLOBAL
41Qwen3.7 MaxALIBABA52.229.540.5%92.3%A$5.40
42MiniMax M3MINIMAX51.829.239.0%92.9%A$0.76SELF-HOST
43Claude Opus 4.6ANTHROPIC51.531.939.9%89.6%A$14.39IN AU
44Qwen3.8 27BALIBABA51.433.733.9%90.5%A$1.62SELF-HOST
45Muse SparkMETA49.831.340.7%88.4%
46Gemini 3 ProGOOGLE48.328.039.7%90.8%A$6.48
47Kimi K2.6MOONSHOT AI46.327.037.5%91.1%A$2.46SELF-HOST
48Claude Sonnet 4.6ANTHROPIC43.630.133.6%87.5%A$8.63IN AU
49Hy3TENCENT40.625.333.5%89.7%A$0.35SELF-HOST
50MiMo V2.5 ProXIAOMI38.926.035.7%86.6%A$0.78SELF-HOST
51GPT-5.4 miniOPENAI33.824.128.1%87.5%A$2.43
52Gemini 3.5 Flash-LiteGOOGLE22.422.218.8%83.8%A$1.22GLOBAL
53Gemma 4 31BGOOGLE18.514.723.6%85.7%A$0.29SELF-HOST
54Mistral Medium 3.5MISTRAL AI0.014.213.8%74.8%A$4.32SELF-HOST
HOW THIS RANKING WORKS

What goes into the reasoning score.

The Intelligence Index already includes Humanity’s Last Exam and GPQA Diamond; this lens weights them up so the ranking leans towards expert-level questions where there is a single right answer.

50%AA Intelligence Index

Artificial Analysis’s composite of its independent evaluations: reasoning, knowledge, maths, coding, long context and agent tasks.

25%Humanity’s Last Exam

Very hard expert-written questions across dozens of fields; frontier models still get most of them wrong.

25%GPQA Diamond

Graduate-level biology, physics and chemistry questions written to be hard to look up.

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 →

QUESTIONS

Choosing a model for reasoning.

As of 2 Oct 2026, Claude Opus 5.5 ranks first, followed by Claude Sonnet 5.5 and Claude Fable 5.1. The ranking combines AA Intelligence Index, Humanity’s Last Exam and GPQA Diamond and is refreshed weekly.

Kimi K3 from Moonshot AI is the highest-ranked open-weights model (14th of 54). Open weights can be run on your own infrastructure, including in an Australian region.

Of the top ten, Muse Spark 1.3 is cheapest at A$2.88 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.

Of the top ten, Claude Opus 5.5 and Claude Opus 5 can run with requests kept in Australia on at least one major cloud platform. See the Australia page for each platform.

The Intelligence Index already includes Humanity’s Last Exam and GPQA Diamond; this lens weights them up so the ranking leans towards expert-level questions where there is a single right answer. Each input is rescaled 0–100 across the models we track, then weighted: AA Intelligence Index 50%, Humanity’s Last Exam 25% and GPQA Diamond 25%.

PUT THE COMPARISON TO WORK

Need help choosing and using AI for your business?