Game Dev

Playable browser games from a single prompt.

Game Dev (HTML, Non-Agentic)

1470

GPT-6 Astra

(max)

1429

Inkling 2 Small

1427

Claude Opus 5.5

1415

Muse Wonder

(max)

1408

GPT-6.1 Sol

1405

Kimi K3.1 Pro

1395

Claude Fable 5.2

1390

Claude Sonnet 5.5

1365

Claude Fable 5.1

1364

GPT-5.6 Sol

(xhigh)

1361

Gemini 4 Argon

1356

Gemini 4 Hydrogen

#175

762

Llama 4 Scout

Evaluates model ability to generate playable browser games as single-file HTML from real-world user prompts. Real-world users interact with the outputs and select their preferred result.

Game Dev (HTML, Non-Agentic) Preference vs Speed

Game Dev (HTML, Non-Agentic) Preference vs Price

Models

GPT-6 Astra (max)

Inkling 2 Small

Muse Ember

MiMo-V2.5

Mercury 3

DeepSeek-V4-Flash-0731

AesCoder-4B

GPT-6 Astra (xhigh)

Claude Opus 5.5

Muse Wonder (max)

GPT-6.1 Sol

Kimi K3.1 Pro

Claude Fable 5.2

Claude Sonnet 5.5

Claude Fable 5.1

GPT-5.6 Sol (xhigh)

Gemini 4 Argon

Gemini 4 Hydrogen

Kimi K3

Muse Spark 1.3 Max

GPT-6.1 Terra

Muse Spark 1.3 (xhigh)

Claude Opus 5

Claude Fable 5

DeepSeek-V4.1-Flash

Gemini 4 Carbon

GPT-5.6 Sol (Medium)

Claude Haiku 5.5

MiMo-V2.6-Pro

Gemini 3.7 Flash

MiniMax M3.1 Flash

Qwen3.8 Max

Ling-3.1-flash

Grok 4.7

GPT-6 Sol

Gemini 3.8 Flash

Muse Spark 1.2

GLM-5.3

GPT-5.5

Grok 4.6

Claude Sonnet 5

Hy4

GLM-5.3-Flash

Claude Opus 4.6 (Thinking)

MiMo-V2.5-Pro

Muse Spark Open (xhigh)

Claude Opus 4.6

Muse Spark 1.1

GLM 5.2

GLM 5.1

Qwen3.7 Max

GPT-6 Luna

Gemini 3.5 Flash

Gemini 3.6 Flash

Grok 4.5

Deca 2.6 Ultra

Deca 2.5 Ultra

Claude Opus 4.8

DeepSeek-V4.1-Flash (1017)

Qwen3.7 Plus

Claude Sonnet 4.6

Deca 2.5 Mini

Deca 2.6 Pro

Kimi K2.6

GLM 5 Turbo

GPT-5.4 (Medium)

GLM 5

DeepSeek-V4-Pro

GPT-5.6 Terra

Claude Opus 4.5

Nex N2 Pro

Kimi K2.7 Code

GLM 5V Turbo

Mistral Large 4 Preview

Qwen3.6 Plus

MiniMax M3

MiniMax M2.7

Hy3

Kimi K2.5 (Thinking)

Gemini 3.1 Pro Preview

GPT-5.6 Luna

DeepSeek-V4-Flash

Gemini 3 Pro Preview

Beam 1.0 (max)

GPT-5.2 (Low)

GPT-5.4 (None)

Grok 4.20 Beta (Reasoning)

GPT-5.4 (Low)

GLM 4.7

GPT-5.2 (High)

GPT-5 (High)

GPT-5.2 (Medium)

GPT-5.1 (High)

Solar Pro 4

MiniMax M2.5

AGI-01 Swift

GPT-5.2 (None)

Grok 4.20 Beta

Grok 4.3

Claude Opus 4.1 (Thinking)

Claude Opus 4

Claude Sonnet 4.5

Gemini 3 Flash Preview

Inkling

Claude Opus 4.1

Claude Sonnet 4.5 (Thinking)

GPT-5.3 Codex

GPT-5.1 (Medium)

GPT-5 (Minimal)

Step 3.7 Flash

GPT-5.1 (Low)

GLM 4.6

GLM 4.5

DeepSeek-V3.2-Exp

Qwen3.5 397B A17B

Claude Sonnet 4 (Thinking)

Claude 3.7 Sonnet

GPT-5.1 Codex

Deca 2.6 Mini

Nemotron 3 Ultra

Hy3 Low

MiniMax M2.1

GLM 4.7 Flash

Claude Sonnet 4

GPT-5.1 (None)

DeepSeek-V3.2

DeepSeek-V3.1 (Thinking)

MiMo-V2-Flash

GPT-5 mini (Default)

Muse Glimmer 30B (xhigh)

MiniMax M2 Stable

Gemini 2.5 Pro

Solar Mini 4

Qwen3.5 Plus 02-15

Qwen3 Coder 480B A35B Instruct

DeepSeek-R1-0528

Claude Haiku 4.5

Qwen3 Max

GLM 4.5 Air

GPT-5.1 Codex Mini

Nemotron 3.5 Super VL

Kimi K2 Turbo Preview

DeepSeek-V3.1

Mistral Large 3 (2512)

GPT-4.1

Mistral Medium 3.5

Trinity Large Thinking

Prime Intellect: INTELLECT-3

Mistral Medium 3.1 (2508)

GPT-4.1 mini

Grok 4.1 Fast

Gemini 2.5 Flash Preview 09-2025

Grok 4 Fast

Grok 4.1 Fast (Reasoning)

DeepSeek-V3-0324

Gemini 2.5 Flash Lite Preview 09-2025

GPT-5 nano (Default)

Ministral 3 14B (2512)

Gemini 3.1 Flash-Lite Preview

Grok 3

Grok 4 Fast (Reasoning)

Mistral Medium 3 (2505)

Gemini 2.5 Flash

o4-mini

Grok 4

Magistral Medium 1.2 (2509)

GPT OSS 120B

Ministral 3 8B (2512)

Grok Code Fast 1

Devstral Medium

Kimi K2

Mercury 2

GPT-4.1 nano

Codestral 2508

Qwen3-235B-A22B-Thinking-2507

Qwen3-235B-A22B-Instruct-2507

Grok 3 Mini

Ministral 3 3B (2512)

Magistral Small 1.2 (2509)

Qwen3-235B-A22B

GPT-4o

Qwen3 30B-A3B

Mistral Small 3.2

Mistral Large 2.1 (2411)

Codestral 2 (2501)

Devstral Small 1.1

The Leaderboard

Updated Nov 13, 2026, 4:06 PM UTC

GPT-6 Astra (max)

1470

800W / 307L

72.3%

±2.6%

1,107

OpenAI

1176.9s

Inkling 2 Small

1429

1343W / 681L

66.3%

±2.1%

2,024

Thinking Machines

661.4s

Claude Opus 5.5

1427

1647W / 939L

63.7%

±1.9%

2,586

Anthropic

490.1s

Muse Wonder (max)

1415

1624W / 990L

62.1%

±1.9%

2,614

Meta

235.7s

GPT-6.1 Sol

1408

1612W / 1070L

60.1%

±1.9%

2,682

OpenAI

225.0s

Kimi K3.1 Pro

1405

1272W / 653L

66.1%

±2.1%

1,925

Moonshot AI

1073.8s

Claude Fable 5.2

1395

362W / 199L

64.5%

±4.0%

561

Anthropic

955.9s

Claude Sonnet 5.5

1390

1243W / 839L

59.7%

±2.1%

2,082

Anthropic

504.0s

Claude Fable 5.1

1365

1075W / 733L

59.5%

±2.3%

1,808

Anthropic

420.7s

#10

GPT-5.6 Sol (xhigh)

1364

1608W / 1001L

61.6%

±1.9%

2,609

OpenAI

341.6s

#11

Gemini 4 Argon

1361

1872W / 1627L

53.5%

±1.7%

3,499

Google

190.3s

#12

Gemini 4 Hydrogen

1356

170W / 129L

56.9%

±5.6%

299

Google

107.7s

#13

Kimi K3

1354

920W / 755L

54.9%

±2.4%

1,675

Moonshot AI

942.6s

#14

Muse Spark 1.3 Max

1345

4682W / 3543L

56.9%

±1.1%

8,225

Meta

299.3s

#15

GPT-6.1 Terra

1334

1351W / 1130L

54.5%

±2.0%

2,481

OpenAI

118.7s

Elo Rating: A skill-based rating system starting at 1200. Ratings are calculated using the standard Bradley-Terry model. See more in our Methodology page.Margin of Error: Shows ±standard error from the Bradley-Terry model for Elo rating confidence intervals.

Game Dev

Playable browser games from a single prompt.

Game Dev (HTML, Non-Agentic)

1470

GPT-6 Astra

(max)

1429

Inkling 2 Small

1427

Claude Opus 5.5

1415

Muse Wonder

(max)

1408

GPT-6.1 Sol

1405

Kimi K3.1 Pro

1395

Claude Fable 5.2

1390

Claude Sonnet 5.5

1365

Claude Fable 5.1

1364

GPT-5.6 Sol

(xhigh)

1361

Gemini 4 Argon

1356

Gemini 4 Hydrogen

#175

762

Llama 4 Scout

Evaluates model ability to generate playable browser games as single-file HTML from real-world user prompts. Real-world users interact with the outputs and select their preferred result.

Game Dev (HTML, Non-Agentic) Preference vs Speed

Game Dev (HTML, Non-Agentic) Preference vs Price

Models

GPT-6 Astra (max)

Inkling 2 Small

Muse Ember

MiMo-V2.5

Mercury 3

DeepSeek-V4-Flash-0731

AesCoder-4B

GPT-6 Astra (xhigh)

Claude Opus 5.5

Muse Wonder (max)

GPT-6.1 Sol

Kimi K3.1 Pro

Claude Fable 5.2

Claude Sonnet 5.5

Claude Fable 5.1

GPT-5.6 Sol (xhigh)

Gemini 4 Argon

Gemini 4 Hydrogen

Kimi K3

Muse Spark 1.3 Max

GPT-6.1 Terra

Muse Spark 1.3 (xhigh)

Claude Opus 5

Claude Fable 5

DeepSeek-V4.1-Flash

Gemini 4 Carbon

GPT-5.6 Sol (Medium)

Claude Haiku 5.5

MiMo-V2.6-Pro

Gemini 3.7 Flash

MiniMax M3.1 Flash

Qwen3.8 Max

Ling-3.1-flash

Grok 4.7

GPT-6 Sol

Gemini 3.8 Flash

Muse Spark 1.2

GLM-5.3

GPT-5.5

Grok 4.6

Claude Sonnet 5

Hy4

GLM-5.3-Flash

Claude Opus 4.6 (Thinking)

MiMo-V2.5-Pro

Muse Spark Open (xhigh)

Claude Opus 4.6

Muse Spark 1.1

GLM 5.2

GLM 5.1

Qwen3.7 Max

GPT-6 Luna

Gemini 3.5 Flash

Gemini 3.6 Flash

Grok 4.5

Deca 2.6 Ultra

Deca 2.5 Ultra

Claude Opus 4.8

DeepSeek-V4.1-Flash (1017)

Qwen3.7 Plus

Claude Sonnet 4.6

Deca 2.5 Mini

Deca 2.6 Pro

Kimi K2.6

GLM 5 Turbo

GPT-5.4 (Medium)

GLM 5

DeepSeek-V4-Pro

GPT-5.6 Terra

Claude Opus 4.5

Nex N2 Pro

Kimi K2.7 Code

GLM 5V Turbo

Mistral Large 4 Preview

Qwen3.6 Plus

MiniMax M3

MiniMax M2.7

Hy3

Kimi K2.5 (Thinking)

Gemini 3.1 Pro Preview

GPT-5.6 Luna

DeepSeek-V4-Flash

Gemini 3 Pro Preview

Beam 1.0 (max)

GPT-5.2 (Low)

GPT-5.4 (None)

Grok 4.20 Beta (Reasoning)

GPT-5.4 (Low)

GLM 4.7

GPT-5.2 (High)

GPT-5 (High)

GPT-5.2 (Medium)

GPT-5.1 (High)

Solar Pro 4

MiniMax M2.5

AGI-01 Swift

GPT-5.2 (None)

Grok 4.20 Beta

Grok 4.3

Claude Opus 4.1 (Thinking)

Claude Opus 4

Claude Sonnet 4.5

Gemini 3 Flash Preview

Inkling

Claude Opus 4.1

Claude Sonnet 4.5 (Thinking)

GPT-5.3 Codex

GPT-5.1 (Medium)

GPT-5 (Minimal)

Step 3.7 Flash

GPT-5.1 (Low)

GLM 4.6

GLM 4.5

DeepSeek-V3.2-Exp

Qwen3.5 397B A17B

Claude Sonnet 4 (Thinking)

Claude 3.7 Sonnet

GPT-5.1 Codex

Deca 2.6 Mini

Nemotron 3 Ultra

Hy3 Low

MiniMax M2.1

GLM 4.7 Flash

Claude Sonnet 4

GPT-5.1 (None)

DeepSeek-V3.2

DeepSeek-V3.1 (Thinking)

MiMo-V2-Flash

GPT-5 mini (Default)

Muse Glimmer 30B (xhigh)

MiniMax M2 Stable

Gemini 2.5 Pro

Solar Mini 4

Qwen3.5 Plus 02-15

Qwen3 Coder 480B A35B Instruct

DeepSeek-R1-0528

Claude Haiku 4.5

Qwen3 Max

GLM 4.5 Air

GPT-5.1 Codex Mini

Nemotron 3.5 Super VL

Kimi K2 Turbo Preview

DeepSeek-V3.1

Mistral Large 3 (2512)

GPT-4.1

Mistral Medium 3.5

Trinity Large Thinking

Prime Intellect: INTELLECT-3

Mistral Medium 3.1 (2508)

GPT-4.1 mini

Grok 4.1 Fast

Gemini 2.5 Flash Preview 09-2025

Grok 4 Fast

Grok 4.1 Fast (Reasoning)

DeepSeek-V3-0324

Gemini 2.5 Flash Lite Preview 09-2025

GPT-5 nano (Default)

Ministral 3 14B (2512)

Gemini 3.1 Flash-Lite Preview

Grok 3

Grok 4 Fast (Reasoning)

Mistral Medium 3 (2505)

Gemini 2.5 Flash

o4-mini

Grok 4

Magistral Medium 1.2 (2509)

GPT OSS 120B

Ministral 3 8B (2512)

Grok Code Fast 1

Devstral Medium

Kimi K2

Mercury 2

GPT-4.1 nano

Codestral 2508

Qwen3-235B-A22B-Thinking-2507

Qwen3-235B-A22B-Instruct-2507

Grok 3 Mini

Ministral 3 3B (2512)

Magistral Small 1.2 (2509)

Qwen3-235B-A22B

GPT-4o

Qwen3 30B-A3B

Mistral Small 3.2

Mistral Large 2.1 (2411)

Codestral 2 (2501)

Devstral Small 1.1

The Leaderboard

Updated Nov 13, 2026, 4:06 PM UTC

GPT-6 Astra (max)

1470

800W / 307L

72.3%

±2.6%

1,107

OpenAI

1176.9s

Inkling 2 Small

1429

1343W / 681L

66.3%

±2.1%

2,024

Thinking Machines

661.4s

Claude Opus 5.5

1427

1647W / 939L

63.7%

±1.9%

2,586

Anthropic

490.1s

Muse Wonder (max)

1415

1624W / 990L

62.1%

±1.9%

2,614

Meta

235.7s

GPT-6.1 Sol

1408

1612W / 1070L

60.1%

±1.9%

2,682

OpenAI

225.0s

Kimi K3.1 Pro

1405

1272W / 653L

66.1%

±2.1%

1,925

Moonshot AI

1073.8s

Claude Fable 5.2

1395

362W / 199L

64.5%

±4.0%

561

Anthropic

955.9s

Claude Sonnet 5.5

1390

1243W / 839L

59.7%

±2.1%

2,082

Anthropic

504.0s

Claude Fable 5.1

1365

1075W / 733L

59.5%

±2.3%

1,808

Anthropic

420.7s

#10

GPT-5.6 Sol (xhigh)

1364

1608W / 1001L

61.6%

±1.9%

2,609

OpenAI

341.6s

#11

Gemini 4 Argon

1361

1872W / 1627L

53.5%

±1.7%

3,499

Google

190.3s

#12

Gemini 4 Hydrogen

1356

170W / 129L

56.9%

±5.6%

299

Google

107.7s

#13

Kimi K3

1354

920W / 755L

54.9%

±2.4%

1,675

Moonshot AI

942.6s

#14

Muse Spark 1.3 Max

1345

4682W / 3543L

56.9%

±1.1%

8,225

Meta

299.3s

#15

GPT-6.1 Terra

1334

1351W / 1130L

54.5%

±2.0%

2,481

OpenAI

118.7s

Overall Preference vs Speed

Overall Preference vs Price

Overall Preference vs Speed

Overall Preference vs Price

Game Dev

Game Dev (HTML, Non-Agentic) Preference vs Speed

Game Dev (HTML, Non-Agentic) Preference vs Price

The Leaderboard

Overall Preference vs Speed

Overall Preference vs Price

Game Dev

Game Dev (HTML, Non-Agentic) Preference vs Speed

Game Dev (HTML, Non-Agentic) Preference vs Price

The Leaderboard