CtrlAltDebrief

CritPt benchmark

381 AI models ranked on CritPt. Each row shows the model’s score next to its cheapest API price per 1M tokens where it has one, so you can weigh quality against cost.

Top model today

Muse Wonder Full (Max) (Meta) leads at 33%

Data via arXiv:2509.26574 →

Models ranked

381

on this benchmark

Top score

33%

current leader

Type

Single eval

benchmark

Updated

Oct 14, 2026

benchmark scores

Cost vs CritPt

247 priced models
Most attractive quadrant (cheap + high)
= reasoning model (extended thinking)
0%10%20%30%40%$0.01$0.03$0.1$0.3$1$3$10$30COST / 1M INPUT TOKENS (log) →CRITPT SCORE ↑GPT-5.6 SolClaude Opus 5.5GPT-6.1 SolGPT-6 AstraClaude Sonnet 5.5GPT-6 SolGPT-5.5 ProGPT-5.6 TerraMiMo-V2.6-ProGLM 5.2DeepSeek V4 Flash 0731DeepSeek V4.1 Flashgpt-oss-20bLlama 3.2 11B Vision Instruct

Each labelled name links to that model’s page. Curated to the cost-efficiency frontier, top scorers and cheapest so labels stay readable; the full field is in the leaderboard table below (and via “select all”). Reasoning models (extended “thinking” before answering) carry a ring around the dot.

Models ranked on CritPt

highest score first · 381 models
Maker
#
1Muse Wonder Full (Max)Meta33%
n/a
n/a
2GPT-5.6 SolOpenAI32%
$2.00
Routed quote
$10.00
Routed quote
3Claude Opus 5.5Anthropic32%
$4.00
via Anthropic
$20.00
via Databricks
4GPT-6.1 SolOpenAI32%
$2.00
via OpenAI
$10.00
via OpenAI
5GPT-6 AstraOpenAI32%
$10.00
via OpenAI
$50.00
via OpenAI
6Claude Sonnet 5.5Anthropic31%
$2.00
via Anthropic
$10.00
via Anthropic
7GPT-6 SolOpenAI31%
$2.00
via OpenAI
$10.00
via OpenAI
8GPT-5.5 ProOpenAI31%
$30.00
via Aihubmix
$180.00
via Aihubmix
9GPT-5.6 TerraOpenAI30%
$2.00
via OpenAI
$12.00
via OpenAI
10GPT-5.4 ProOpenAI30%
$30.00
Routed quote
$180.00
Routed quote
11Claude Fable 5.1Anthropic30%
$10.00
via Anthropic
$50.00
via Anthropic
12Claude Opus 5Anthropic29%
$5.00
via Anthropic
$25.00
via Anthropic
13Claude Fable 5Anthropic29%
$10.00
via Anthropic
$50.00
via Anthropic
14Muse WonderOther28%
$4.00
Routed quote
$20.00
Routed quote
15GPT-5.5OpenAI27%
$5.00
via OpenAI
$30.00
via OpenAI
16Gemini 4 ArgonGoogle27%
$2.00
via Google
$10.00
via Google
17MiMo-V2.6-ProXiaomi27%
$0.435
via Xiaomi Mimo
$0.87
via Xiaomi Mimo
18Gemini 3 Deep ThinkGoogle26%
n/a
n/a
19Muse Spark 1.3Other25%
$1.25
via Azure
$4.25
via Azure
20GPT-5.4OpenAI23%
$2.50
via Databricks
$15.00
via OpenAI
21Kimi K3Moonshot23%
$0.70
Routed quote
$12.50
via Sail
22GLM 5.2Zhipu21%
$0.05
Routed quote
$1.98
via Scx AI
23Claude Opus 4.8Anthropic21%
$5.00
via Anthropic
$25.00
via Anthropic
24Step 5 PreviewStepfun21%
n/a
n/a
25GPT-5.6 LunaOpenAI21%
$0.20
via OpenAI
$1.20
via OpenAI
26Qwen3.8 MaxAlibaba20%
$1.65
via DeepInfra
$4.95
via DeepInfra
27Qwen3.8 2.4T A95BAlibaba20%
$2.00
via Aihubmix
$6.00
via Aihubmix
28GPT-6 LunaOpenAI19%
$0.10
via OpenAI
$0.50
via OpenAI
29GLM 5.3Zhipu19%
$0.079
Routed quote
$3.08
via Sail
30Claude Haiku 5.5Anthropic19%
$0.10
via Anthropic
$0.50
via Anthropic
31Gemini 3.8 FlashGoogle18%
$0.75
via Google
$3.75
via Google
32DeepSeek V4 Pro 0813DeepSeek18%
$0.66
Routed quote
$1.98
Routed quote
33Ling 3.1 FlashInclusionAI18%
n/a
n/a
34Gemini 3.1 Pro PreviewGoogle18%
$2.00
via Google
$12.00
via Google
35Grok 4.7xAI18%
$2.00
via xAI
$6.00
via xAI
36Muse Spark 1.2Other18%
$1.25
via Meta
$4.25
via Meta
37Qwen3.8 Max (0902)Alibaba18%
$2.00
Routed quote
$6.00
Routed quote
38Grok 4.6xAI17%
$2.00
via xAI
$6.00
via xAI
39Claude Sonnet 5Anthropic17%
$2.00
via Anthropic
$10.00
via Anthropic
40GPT-5.3-CodexOpenAI17%
$1.75
via Aihubmix
$14.00
via Aihubmix
41DeepSeek V4 Flash 0731DeepSeek17%
$0.016
Routed quote
$0.153
Routed quote
42Agnes 2.5 Pro BetaSapiens-ai16%
n/a
n/a
43GLM 5.3 FlashZhipu15%
$0.11
via Sail
$0.35
via Sail
44Grok 4.5xAI15%
$2.00
via xAI
$6.00
via xAI
45Muse Spark 1.1Other15%
$1.25
via Meta
$4.25
via Meta
46DeepSeek V4.1 FlashDeepSeek14%
$0.01
via Coralbricks
$0.60
via Sail
47Gemini 3.7 FlashGoogle14%
$0.75
via Google
$3.75
via Google
48JT-4.1 Flash 236B A21BChina-mobile14%
n/a
n/a
49JT-4.1 Flash 236B A21BChina-mobile14%
n/a
n/a
50Qwen3.7 MaxAlibaba13%
$1.25
via Novita
$3.75
via Novita
1–50 of 389
Page 1 of 8
On-demand listed rates, OpenRouter routed quotes and Vercel reseller quotes. Enterprise commitments can be materially lower. Cost per task uses best-effort token usage. An OpenRouter routed quote is the price of whichever upstream provider OpenRouter prefers today, so it moves day to day.

Frequently asked questions

What is the CritPt benchmark?

CritPt (Complex Research using Integrated Thinking, Physics Test) is a frontier physics benchmark of 71 unpublished, research-level challenges across 12 subdisciplines, created by more than 50 active physicists. Scored on accuracy in guess-resistant formats (numeric arrays, symbolic expressions, code), where the best current models solve only single-digit percentages unaided.

Which AI model scores highest on CritPt?

Muse Wonder Full (Max) (Meta) leads with 33%.

How many models are ranked on CritPt?

381 models carry a CritPt score. Scores come from Artificial Analysis; where a model has a listed API price it is drawn from LiteLLM and OpenRouter, and models with no current price are listed without one.

Other benchmarks

compare the same models on a different eval
Intelligence Indexcomposite indexCoding Indexcomposite indexAgentic Indexcomposite indexGPQA DiamondleaderboardHumanity's Last ExamleaderboardMMLU-ProleaderboardMMMU-ProleaderboardLiveCodeBenchleaderboardSciCodeleaderboardTerminalBench HardleaderboardTerminalBench v2.1leaderboardτ²-benchleaderboardτ-bench BankingleaderboardAPEX AgentsleaderboardIT-Bench SREleaderboardGDPvalleaderboardOmniscienceleaderboardIFBenchleaderboardLong-Context Reasoningcomposite indexSWE-bench VerifiedleaderboardSWE-bench ProleaderboardAIME 2026leaderboardARC-AGI-2leaderboardAA-Briefcaseleaderboard
Benchmark scores via Artificial Analysis · scores updated Oct 14, 2026 · Listed prices from LiteLLM, OpenRouter and Vercel · catalogue refreshed Oct 14, 2026 · 148 prices older, back to Jul 19, 2026

Every weekday

AI moves fast. Here's your debrief.

News, analysis, tools, and more.

For people who build with AI