Best value AI models.

Capability per Australian dollar: independent test results weighed against list price, with the models no one beats on both marked out. As of 23 Sep 2026, GLM-5.3 Flash leads for value, ahead of MiMo V2.6 Pro and Qwen3.8 Flash Next, out of 50 ranked models. GLM-5.3 Flash has open weights, so it can also be self-hosted.

50 MODELS RANKEDA$1 = US$0.7123UPDATED 23 SEP 2026
CAPABILITY AGAINST PRICEHOVER OR TAP A DOT
Intelligence Index against price per million tokens in AUD for 50 models; 6 on the value frontier.102030405060A$0.3A$1A$3A$10A$30A$ PER MILLION TOKENS (LOG SCALE) →↑ INTELLIGENCE INDEXGPT-6 LunaQwen3.8 Flash NextGLM-5.3 FlashMiMo V2.6 ProMuse Spark 1.3Claude Opus 5.5GPT-6 AstraGrok 4.7Qwen3.8 MaxStep 5Kimi K3Gemini 3.8 FlashDeepSeek V4.1 FlashMiniMax M3Hy3Mistral Medium 3.5
Value frontier — no tracked model is both smarter and cheaper Other models — each lab’s best is named where it fitsHover or tap a dot for details · pick a lab to name all its models · swipe the chart sideways to see it all
VALUE RANKING50 MODELS · 23 SEP 2026
50 MODELS
Best value AI models, ranked, as of 23 Sep 2026
MODELFRONTIERIN AUSTRALIA
01GLM-5.3 FlashZ.AI76.741.81475YESA$0.33SELF-HOST
02MiMo V2.6 ProXIAOMI75.746.3YESA$0.76NOT IN AU
03Qwen3.8 Flash NextALIBABA74.239.8YESA$0.32SELF-HOST
04GPT-6 LunaOPENAI71.937.3YESA$0.28GLOBAL
05Claude Opus 5.5ANTHROPIC68.057.6YESA$11.23IN AU
06Muse Spark 1.3META66.948.11493YESA$2.81NOT IN AU
07DeepSeek V4.1 FlashDEEPSEEK66.639.5A$0.74SELF-HOST
08GPT-5.6 LunaOPENAI64.937.31453A$0.63GLOBAL
09Step 5STEPFUN63.743.7A$2.00
10GLM-5.3Z.AI61.744.81483A$3.02SELF-HOST
11Grok 4.7XAI61.046.4A$4.21NOT IN AU
12GPT-6 SolOPENAI60.047.5A$5.62GLOBAL
13Qwen3.8 MaxALIBABA59.645.41481A$4.21NOT IN AU
14Gemini 3.8 FlashGOOGLE59.440.91493A$2.11GLOBAL
15Grok 4.6XAI58.144.31456A$4.21GLOBAL
16DeepSeek V4 FlashDEEPSEEK57.434.3A$0.93GLOBAL
17Gemini 3.7 FlashGOOGLE56.939.11490A$2.11GLOBAL
18Claude Opus 5ANTHROPIC56.650.81493A$14.04IN AU
19Muse Spark 1.2META55.139.61500A$2.81NOT IN AU
20Claude Fable 5.1ANTHROPIC54.253.41499A$28.08GLOBAL
21Hy3TENCENT53.425.31456A$0.35SELF-HOST
22GPT-5.6 SolOPENAI53.347.01484A$11.23GLOBAL
23GPT-6 AstraOPENAI53.252.71480A$28.08GLOBAL
24MiniMax M3MINIMAX52.429.21441A$0.74SELF-HOST
25Qwen3.8 27BALIBABA52.033.71437A$1.58SELF-HOST
26GPT-5.6 TerraOPENAI51.542.11466A$6.32GLOBAL
27Kimi K3MOONSHOT AI51.143.61485A$8.42GLOBAL
28Grok 4.5XAI50.538.81468A$4.21NOT IN AU
29DeepSeek V4 ProDEEPSEEK50.236.01463A$2.78GLOBAL
30Gemini 3.6 FlashGOOGLE49.934.01480A$2.11GLOBAL
31Claude Fable 5ANTHROPIC48.949.61506A$28.08GLOBAL
32MiMo V2.5 ProXIAOMI47.626.01467A$0.76SELF-HOST
33Claude Sonnet 5ANTHROPIC47.238.21461A$5.62IN AU
34Muse Spark 1.1META47.033.71493A$2.81
35Gemma 4 31BGOOGLE46.419.01451A$0.29SELF-HOST
36GLM-5.2Z.AI46.333.71472A$3.02SELF-HOST
37GPT-5.4OPENAI45.339.01476A$7.90IN AU
38Claude Opus 4.8ANTHROPIC44.241.81481A$14.04IN AU
39Claude Opus 4.7ANTHROPIC42.740.71502A$14.04IN AU
40Gemini 3.5 FlashGOOGLE40.932.61478A$4.74IN AU
41Kimi K2.6MOONSHOT AI39.027.01460A$2.40SELF-HOST
42Gemini 3.5 Flash-LiteGOOGLE38.522.21456A$1.19GLOBAL
43GPT-5.5OPENAI38.538.41482A$15.79GLOBAL
44Qwen3.7 MaxALIBABA35.729.51473A$5.26
45GPT-5.4 miniOPENAI35.224.11448A$2.37
46Gemini 3.1 ProGOOGLE34.429.71487A$6.32GLOBAL
47Claude Sonnet 4.6ANTHROPIC32.430.11473A$8.42IN AU
48Gemini 3 ProGOOGLE32.028.01486A$6.32
49Claude Opus 4.6ANTHROPIC30.531.91505A$14.04IN AU
50Mistral Medium 3.5MISTRAL AI16.514.21426A$4.21SELF-HOST
HOW THIS RANKING WORKS

What goes into the value score.

Quality is the Intelligence Index; price is the blended AUD list price on a log scale, so going from A$30 to A$3 per million tokens counts the same as A$3 to A$0.30. A model is marked “frontier” when no tracked model is both smarter and cheaper.

60%AA Intelligence Index

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

40%Blended price (AUD)

List price per million tokens, blended 3 parts input to 1 part output, converted to Australian dollars at the snapshot’s RBA rate. Excludes GST.

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 value.

As of 23 Sep 2026, GLM-5.3 Flash ranks first, followed by MiMo V2.6 Pro and Qwen3.8 Flash Next. The ranking combines AA Intelligence Index and Blended price (AUD) and is refreshed weekly.

GLM-5.3 Flash from Z.ai is the highest-ranked open-weights model (1st of 50). Open weights can be run on your own infrastructure, including in an Australian region.

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.

Quality is the Intelligence Index; price is the blended AUD list price on a log scale, so going from A$30 to A$3 per million tokens counts the same as A$3 to A$0.30. A model is marked “frontier” when no tracked model is both smarter and cheaper. Each input is rescaled 0–100 across the models we track, then weighted: AA Intelligence Index 60% and Blended price (AUD) 40%.

PUT THE COMPARISON TO WORK

Need help choosing and using AI for your business?