Skip to content
llmwaves

AI trends

What actually changed

Two things move at once: the ceiling keeps rising, and the floor price of any given capability keeps collapsing. The second one changes more decisions than the first.

Frontier record holder

63.1

Claude Opus 5 · Jul 2026

Index points gained

+56.3

since May 2023

Price of Index 40+

44× cheaper

since it first appeared

Frontier records set

18

distinct record holders

The ceiling

Frontier intelligence over time

Best Intelligence Index score available on any given date. Flat stretches are the interesting part — that is how long a record held.

Frontier Intelligence Index

Single series · a step means a new record holder

0102030405060202420252026Intelligence Index

Hover the plot for the model holding the record at that date.

View as table
DateModelIndex
May 2023GPT-4 (OpenAI)6.8
Apr 2024GPT-4 Turbo (OpenAI)7.7
May 2024GPT-4o (OpenAI)11.1
Dec 2024o1 (OpenAI)23.9
Apr 2025o4 Mini (OpenAI)26.1
Apr 2025o3 (OpenAI)31.1
Jun 2025o3 Pro (OpenAI)33.3
Aug 2025GPT-5 (OpenAI)35.3
Nov 2025GPT-5.1-Codex (OpenAI)35.6
Nov 2025GPT-5.1 (OpenAI)37.5
Dec 2025GPT-5.2 (OpenAI)43.3
Feb 2026Gemini 3.1 Pro Preview (Google)47.7
Mar 2026GPT-5.4 (OpenAI)53.1
Apr 2026Claude Opus 4.7 (Anthropic)55.0
Apr 2026GPT-5.5 (OpenAI)56.3
May 2026Claude Opus 4.8 (Anthropic)57.3
Jun 2026Claude Fable 5 (Anthropic)62.1
Jul 2026Claude Opus 5 (Anthropic)63.1

The floor

What a capability costs over time

Cheapest blended price at which each intelligence tier was available. Each line steps down when somebody undercuts the previous best.

Cheapest price per intelligence tier

Blended USD per 1M tokens on a log scale · lower is better

  • Index 30+
  • Index 40+
  • Index 55+
$0.01$0.1$1$10JulOctJan 2026AprJulBlended USD / 1M tokensIndex 55+Index 40+Index 30+

A tier only appears once some model reached it. Log scale, because the drops are order-of-magnitude. Where one model is the cheapest option for two tiers at once, those lines coincide — hover to read every tier at a date.

View as table
TierDateModelBlended / 1M
Index 30+Apr 2025o3$3.5
Index 30+Aug 2025GPT-5$3.44
Index 30+Nov 2025Kimi K2 Thinking$1.07
Index 30+Nov 2025GPT-5.1-Codex-Mini$0.688
Index 30+Dec 2025M2.1$0.525
Index 30+Feb 2026M2.5$0.39
Index 30+Apr 2026MiMo-V2.5$0.175
Index 30+Apr 2026Hy3 preview$0.1
Index 30+Jul 2026Ling-3.0-flash$0.032
Index 40+Dec 2025GPT-5.2$4.81
Index 40+Feb 2026GLM 5$1.35
Index 40+Apr 2026Qwen3.6 Plus$0.731
Index 40+Apr 2026MiMo-V2.5-Pro$0.544
Index 40+Apr 2026V4 Flash 0423$0.11
Index 55+Apr 2026Claude Opus 4.7$10
Index 55+Jun 2026Claude Sonnet 5$4
Index 55+Jul 2026Grok 4.5$3
Index 55+Jul 2026GPT-5.6 Terra$2.25
Index 55+Aug 2026Muse Spark 1.2$2

The race

Six creators, best-so-far

Each line is one creator's own record over time. Convergence at the right edge is the story of the last year.

Best Intelligence Index per creator

Six series · each line steps when that creator beats its own record

  • Anthropic
  • OpenAI
  • Google
  • Qwen
  • DeepSeek
  • Meta
0102030405060202420252026Intelligence IndexAnthropicOpenAIQwenGoogleDeepSeekMeta

Only creators with more than one benchmarked release appear. Lines are direct-labelled at the right edge.

View as table
CreatorDateModelIndex
AnthropicMar 2024Claude 3 Haiku3.5
AnthropicMay 2025Claude Sonnet 426.0
AnthropicAug 2025Claude Opus 4.128.8
AnthropicSep 2025Claude Sonnet 4.529.9
AnthropicNov 2025Claude Opus 4.535.6
AnthropicFeb 2026Claude Opus 4.638.8
AnthropicApr 2026Claude Opus 4.755.0
AnthropicMay 2026Claude Opus 4.857.3
AnthropicJun 2026Claude Fable 562.1
AnthropicJul 2026Claude Opus 563.1
OpenAIMay 2023GPT-46.8
OpenAIApr 2024GPT-4 Turbo7.7
OpenAIMay 2024GPT-4o11.1
OpenAIDec 2024o123.9
OpenAIApr 2025o4 Mini26.1
OpenAIApr 2025o331.1
OpenAIJun 2025o3 Pro33.3
OpenAIAug 2025GPT-535.3
OpenAINov 2025GPT-5.1-Codex35.6
OpenAINov 2025GPT-5.137.5
OpenAIDec 2025GPT-5.243.3
OpenAIFeb 2026GPT-5.3-Codex45.5
OpenAIMar 2026GPT-5.453.1
OpenAIApr 2026GPT-5.556.3
OpenAIJul 2026GPT-5.6 Sol60.9
GoogleJun 2025Gemini 2.5 Pro25.9
GoogleFeb 2026Gemini 3.1 Pro Preview47.7
GoogleMay 2026Gemini 3.5 Flash52.0
QwenSep 2024Qwen2.5 72B Instruct9.4
QwenJul 2025Qwen3 Coder 30B A3B Instruct13.6
QwenSep 2025Qwen3 Next 80B A3B Instruct13.8
QwenSep 2025Qwen3 Max24.5
QwenFeb 2026Qwen3 Max Thinking32.5
QwenFeb 2026Qwen3.5 397B A17B34.3
QwenFeb 2026Qwen3.5-27B34.6
QwenApr 2026Qwen3.6 Plus40.5
QwenApr 2026Qwen3.6 Max Preview41.1
QwenMay 2026Qwen3.7 Max46.7
QwenAug 2026Qwen3.8 Max58.1
DeepSeekJan 2025R120.4
DeepSeekSep 2025V3.1 Terminus21.7
DeepSeekDec 2025V3.225.1
DeepSeekApr 2026V4 Flash 042351.8
MetaApr 2025Llama 4 Scout10.3
MetaApr 2025Llama 4 Maverick14.5

Cadence

Releases and standings

How many models shipped per quarter, and where each creator's best model currently sits.

Models released per quarter

Count of models entering the index · last twelve quarters

  • 2023-Q3
    5
  • 2024-Q1
    4
  • 2024-Q2
    5
  • 2024-Q3
    17
  • 2024-Q4
    14
  • 2025-Q1
    24
  • 2025-Q2
    28
  • 2025-Q3
    46
  • 2025-Q4
    44
  • 2026-Q1
    47
  • 2026-Q2
    56
  • 2026-Q3
    31

Counts the models tracked here, not every model released anywhere.

View as table
QuarterModels
2023-Q35
2024-Q14
2024-Q25
2024-Q317
2024-Q414
2025-Q124
2025-Q228
2025-Q346
2025-Q444
2026-Q147
2026-Q256
2026-Q331

Best model per creator

Intelligence Index of each creator's strongest release

One entry per creator — this is a standings table, not a model ranking.

View as table
CreatorBest modelIndex
AnthropicClaude Opus 563.1
OpenAIGPT-5.6 Sol60.9
Moonshot AIKimi K359.7
QwenQwen3.8 Max58.1
MetaMuse Spark 1.256.8
xAIGrok 4.555.8
Z.aiGLM 5.252.6
GoogleGemini 3.5 Flash52.0
DeepSeekV4 Flash 073151.8
MiniMaxMiniMax M345.4
XiaomiMiMo-V2.5-Pro42.9
Thinking MachinesInkling42.3