CtrlAltDebrief

SciCode benchmark

386 AI models ranked on SciCode. 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

Claude Fable 5 (Anthropic) leads at 60%, from $10.00 per 1M input.

Data via arXiv:2407.13168 →

Models ranked

386

on this benchmark

Top score

60%

current leader

Type

Single eval

benchmark

Updated

15 July 2026

last source fetch

Cost vs SciCode

204 priced models
Most attractive quadrant (cheap + high)
= reasoning model (extended thinking)
0%20%40%60%$0.01$0.03$0.1$0.3$1$3$10COST / 1M INPUT TOKENS (log) →SCICODE SCORE ↑Claude Fable 5Gemini 3.1 Pro PreviewMuse Spark 1.1GPT-5.6 SolGPT-5.4GPT-5.5Gemini 3 ProGPT-5.2-CodexGPT-5.6 LunaDeepSeek V4 ProMiniMax M2.7GPT-5.4 NanoGemma 4 31BQwen3 235B A22b Instruct 2507gpt-oss-120bGPT-5 Nanogpt-oss-20bQwen2 5 Coder 7B InstructLlama 3.1 8B 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 SciCode

highest score first · 386 models
Maker
#
1Claude Fable 5Anthropic60%$10.00$50.00
2Gemini 3.1 Pro PreviewGoogle59%$2.00$12.00
3Muse Spark 1.1Other58%$1.25$4.25
4GPT-5.6 SolOpenAI57%high$5.00$30.00
5GPT-5.4OpenAI57%$2.50$15.00
6GPT-5.5OpenAI56%$5.00$30.00
7Gemini 3 ProGoogle56%$2.00$12.00
8GPT-5.2-CodexOpenAI55%$1.75$14.00
9Claude Opus 4.7Anthropic55%$5.00$25.00
10Grok 4.5xAI54%$2.00$6.00
11GPT-5.6 TerraOpenAI54%$2.50$15.00
12Claude Sonnet 5Anthropic54%$2.00$10.00
13Claude Opus 4.8Anthropic53%$5.00$25.00
14Kimi K2.6Moonshot53%n/an/a
15GPT-5.3 CodexOpenAI53%n/an/a
16Gemini 3.5 FlashGoogle53%$1.50$9.00
17GPT-5.6 LunaOpenAI53%$1.00$6.00
18GPT-5.2OpenAI52%$1.75$14.00
19Claude Opus 4.6Anthropic52%reasoning: adaptive$5.00$25.00
20Muse SparkMeta52%n/an/a
21Gemini 3 FlashGoogle51%reasoning: truen/an/a
22GLM 5.2Zhipu50%$1.40$4.40
23GPT-5.5 Instant (May 2026)OpenAI50%n/an/a
24MiMo-V2.5-ProXiaomi50%$1.00$3.00
25Grok Build 0.1 0616Spacexai50%n/an/a
26DeepSeek V4 ProDeepSeek50%$0.43$0.87
27GPT-5.4 MiniOpenAI50%$0.75$4.50
28Claude Opus 4.5Anthropic50%reasoning: true$5.00$25.00
29Kimi K2.5Moonshot49%$0.50$2.80
30Qwen3.7 MaxAlibaba49%n/an/a
31GPT-5.5 Instant (June 2026)OpenAI49%n/an/a
32Hy3-previewTencent48%n/an/a
33Kimi K2.7 CodeMoonshot47%$0.95$4.00
34Grok 4.3xAI47%$1.25$2.50
35MiniMax M2.7MiniMax47%$0.26$0.55
36Claude Sonnet 4.6Anthropic47%$3.00$15.00
37GPT-5.4 NanoOpenAI47%$0.20$1.25
38Qwen3.6 Max PreviewAlibaba47%n/an/a
39o4 MiniOpenAI47%$1.00$4.00
40GLM 5Zhipu46%$0.80$2.56
41Grok 4xAI46%$3.00$15.00
42Grok 4.20 0309 v2Spacexai46%n/an/a
43Qwen3.7 PlusAlibaba45%n/an/a
44MiniMax M3MiniMax45%$0.30$1.20
45GLM 4.7Zhipu45%$0.40$1.50
46DeepSeek V4 FlashDeepSeek45%n/an/a
47Claude 4.5 SonnetAnthropic45%reasoning: true$3.00$15.00
48Grok 4.20 0309Spacexai45%n/an/a
49Grok 4.1 Fast ReasoningxAI44%$0.20$0.50
50Grok 4 Fast ReasoningxAI44%$0.20$0.50
1–50 of 386
Page 1 of 8
Prices are on-demand list rates. Enterprise commitments can be materially lower. Cost per task uses best-effort token usage.

Frequently asked questions

What is the SciCode benchmark?

A research coding benchmark curated by scientists across 16 natural-science sub-fields, where models write Python to solve real scientific problems. Its 80 main problems decompose into 338 subproblems, graded against scientist-annotated gold solutions and test cases.

Which AI model scores highest on SciCode?

Claude Fable 5 (Anthropic) leads with 60%, from $10.00 per 1M input tokens.

How many models are ranked on SciCode?

386 models carry a SciCode score. Scores come from Artificial Analysis; where a model has a live API price it is verified against 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-ProleaderboardLiveCodeBenchleaderboardTerminalBench HardleaderboardTerminalBench v2.1leaderboardτ²-benchleaderboardτ-bench BankingleaderboardAPEX AgentsleaderboardIT-Bench SREleaderboardGDPvalleaderboardCritPtleaderboardOmniscienceleaderboardIFBenchleaderboardLong-Context Reasoningcomposite indexSWE-bench VerifiedleaderboardSWE-bench ProleaderboardAIME 2026leaderboardARC-AGI-2leaderboardAA-Briefcaseleaderboard
Benchmark scores via Artificial Analysis · pricing verified against LiteLLM + OpenRouter · refreshed daily · updated 15 July 2026

Every weekday

AI moves fast. Here's your debrief.

News, analysis, tools, and more.

For people who build with AI