New: Agent Index for knowledge work is live.Explore the index ↗
Compare models

Best LLMs for knowledge work

GDPval-AA

GDPval-AA evaluates models on professional knowledge-work tasks from OpenAI’s GDPval dataset, including creating documents, spreadsheets, slides, and other work products. Agents use tools through the Stirrup framework. Results are Elo ratings from blind pairwise comparisons, rather than percentage accuracy.

Updated: Data as of:
Widely usedSaturation: not applicable
Official benchmark
GDPval-AA results. Each row identifies its benchmark version and evaluation track. Scores from different versions are not directly comparable. Best effort from the latest stored report date per model and version.
Model / creatorVersionReported score
Claude Opus 5.5Anthropic · Proprietary2.11,866.27 Elo
Kimi K3.1 ProMoonshot AI · Open source / open weights2.11,858.05 Elo
Claude Sonnet 5.5Anthropic · Proprietary2.11,837.82 Elo
Claude Fable 5.1Anthropic · Proprietary2.11,763.29 Elo
DeepSeek V4.1 Flash 1017DeepSeek · Open source / open weights2.11,745.77 Elo
Gemini 4 CarbonGoogle · Proprietary2.11,732.68 Elo
Claude Opus 5Anthropic · Proprietary2.11,729.9 Elo
Grok 4.7xAI · Proprietary2.11,718.42 Elo
Muse Spark 1.3Meta · Proprietary2.11,680.99 Elo
GLM-5.3Z.ai · Open source / open weights2.11,656.88 Elo
GLM-5.3-FlashZ.ai · Open source / open weights2.11,649.37 Elo
Gemini 4 ArgonGoogle · Proprietary2.11,630.66 Elo
Claude Fable 5Anthropic · Proprietary2.11,618.39 Elo
GPT-5.6 SolOpenAI · Proprietary2.11,617.83 Elo
GPT-6.1 SolOpenAI · Proprietary2.11,608.89 Elo
DeepSeek V4.1 FlashDeepSeek · Open source / open weights2.11,600 Elo
GPT-6 AstraOpenAI · Proprietary2.11,583.69 Elo
Kimi K3Moonshot AI · Open source / open weights2.11,542.76 Elo
Muse Spark 1.2Meta · Proprietary2.11,480.04 Elo
Claude Sonnet 5Anthropic · Proprietary2.11,470.38 Elo
MiniMax M3.1 FlashMiniMax · Open source / open weights2.11,465.36 Elo
Claude Opus 4.8Anthropic · Proprietary2.11,461.75 Elo
DeepSeek V4 Pro 0813DeepSeek · Open source / open weights2.11,457.37 Elo
Gemini 3.8 FlashGoogle · Proprietary2.11,445.43 Elo
GLM-5.2Z.ai · Open source / open weights2.11,381.49 Elo
Claude Opus 4.7Anthropic · Proprietary2.11,361.69 Elo
GPT-5.5OpenAI · Proprietary2.11,359.24 Elo
MiniMax M3MiniMax · Open source / open weights2.11,249.95 Elo
Muse Spark 1.1Meta · Proprietary2.11,217.25 Elo
DeepSeek V4 Pro 0424DeepSeek · Open source / open weights2.11,166.24 Elo
Kimi K2.6Moonshot AI · Open source / open weights2.11,046.53 Elo
DeepSeek V4 Flash 0420DeepSeek · Open source / open weights2.11,045.62 Elo
Muse SparkMeta · Proprietary2.11,006.05 Elo
Kimi K2.5Moonshot AI · Open source / open weights2.1846.64 Elo
Gemini 3.1 ProGoogle · Proprietary2.1798.88 Elo
gpt-oss-120bOpenAI · Open source / open weights2.1615.64 Elo
gpt-oss-20bOpenAI · Open source / open weights2.1347.19 Elo
DeepSeek R1 (Jan 2025)DeepSeek · Open source / open weights2.1302.47 Elo
View all 89 published reports
Claude Opus 5.51,866.27 Elo
2.1

Claude Opus 5.5 (Max, Default Fallback); 220 tasks; Stirrup agent; pairwise Elo

Kimi K3.1 Pro1,858.05 Elo
2.1

Kimi K3.1 Pro; 220 tasks; Stirrup agent; pairwise Elo

Claude Sonnet 5.51,837.82 Elo
2.1

Claude Sonnet 5.5 (Max, Default Fallback); 220 tasks; Stirrup agent; pairwise Elo

Claude Opus 5.51,835.03 Elo
2.1

Claude Opus 5.5 (Xhigh, Default Fallback); 220 tasks; Stirrup agent; pairwise Elo

Claude Fable 5.11,763.29 Elo
2.1

Claude Fable 5.1 (Max, Default Fallback); 220 tasks; Stirrup agent; pairwise Elo

Claude Fable 5.11,740.22 Elo
2.1

Claude Fable 5.1 (Xhigh, Default Fallback); 220 tasks; Stirrup agent; pairwise Elo

Claude Sonnet 5.51,735.08 Elo
2.1

Claude Sonnet 5.5 (Xhigh, Default Fallback); 220 tasks; Stirrup agent; pairwise Elo

Gemini 4 Carbon1,732.68 Elo
2.1

Gemini 4 Carbon (High); 220 tasks; Stirrup agent; pairwise Elo

Claude Opus 51,729.9 Elo
2.1

Claude Opus 5 (Max); 220 tasks; Stirrup agent; pairwise Elo

Grok 4.71,718.42 Elo
2.1

Grok 4.7 (Xhigh); 220 tasks; Stirrup agent; pairwise Elo

Grok 4.71,715.8 Elo
2.1

Grok 4.7 (High); 220 tasks; Stirrup agent; pairwise Elo

Claude Opus 5.51,713.6 Elo
2.1

Claude Opus 5.5 (High, Default Fallback); 220 tasks; Stirrup agent; pairwise Elo

Claude Opus 51,695.66 Elo
2.1

Claude Opus 5 (Xhigh); 220 tasks; Stirrup agent; pairwise Elo

Muse Spark 1.31,680.99 Elo
2.1

Muse Spark 1.3 (Max); 220 tasks; Stirrup agent; pairwise Elo

GLM-5.31,656.88 Elo
2.1

GLM-5.3 (Max); 220 tasks; Stirrup agent; pairwise Elo

GLM-5.3-Flash1,649.37 Elo
2.1

GLM-5.3-Flash; 220 tasks; Stirrup agent; pairwise Elo

Claude Fable 5.11,644.08 Elo
2.1

Claude Fable 5.1 (High, Default Fallback); 220 tasks; Stirrup agent; pairwise Elo

Muse Spark 1.31,631.96 Elo
2.1

Muse Spark 1.3 (Xhigh); 220 tasks; Stirrup agent; pairwise Elo

Gemini 4 Argon1,630.66 Elo
2.1

Gemini 4 Argon (High); 220 tasks; Stirrup agent; pairwise Elo

Claude Fable 51,618.39 Elo
2.1

Claude Fable 5 (Max, Opus 4.8 Fallback); 220 tasks; Stirrup agent; pairwise Elo

GPT-5.6 Sol1,617.83 Elo
2.1

GPT-5.6 Sol (Max); 220 tasks; Stirrup agent; pairwise Elo

GPT-6.1 Sol1,608.89 Elo
2.1

GPT-6.1 Sol (Max); 220 tasks; Stirrup agent; pairwise Elo

Claude Opus 51,603.96 Elo
2.1

Claude Opus 5 (High); 220 tasks; Stirrup agent; pairwise Elo

2.1

DeepSeek V4.1 Flash (Max); 220 tasks; Stirrup agent; pairwise Elo

Claude Opus 5.51,592.56 Elo
2.1

Claude Opus 5.5 (Medium, Default Fallback); 220 tasks; Stirrup agent; pairwise Elo

Grok 4.71,591.4 Elo
2.1

Grok 4.7 (Low); 220 tasks; Stirrup agent; pairwise Elo

GPT-6 Astra1,583.69 Elo
2.1

GPT-6 Astra (Max); 220 tasks; Stirrup agent; pairwise Elo

GPT-5.6 Sol1,577.77 Elo
2.1

GPT-5.6 Sol (Xhigh); 220 tasks; Stirrup agent; pairwise Elo

GPT-6 Astra1,569.36 Elo
2.1

GPT-6 Astra (Xhigh); 220 tasks; Stirrup agent; pairwise Elo

Claude Sonnet 5.51,556.89 Elo
2.1

Claude Sonnet 5.5 (High, Default Fallback); 220 tasks; Stirrup agent; pairwise Elo

GPT-6.1 Sol1,554.8 Elo
2.1

GPT-6.1 Sol (Xhigh); 220 tasks; Stirrup agent; pairwise Elo

Claude Fable 5.11,554.23 Elo
2.1

Claude Fable 5.1 (Medium, Default Fallback); 220 tasks; Stirrup agent; pairwise Elo

Kimi K31,542.76 Elo
2.1

Kimi K3 (Max); 220 tasks; Stirrup agent; pairwise Elo

GPT-6 Astra1,536.04 Elo
2.1

GPT-6 Astra (High); 220 tasks; Stirrup agent; pairwise Elo

GPT-6.1 Sol1,520.45 Elo
2.1

GPT-6.1 Sol (High); 220 tasks; Stirrup agent; pairwise Elo

GPT-5.6 Sol1,511.8 Elo
2.1

GPT-5.6 Sol (High); 220 tasks; Stirrup agent; pairwise Elo

Claude Opus 51,498.64 Elo
2.1

Claude Opus 5 (Medium); 220 tasks; Stirrup agent; pairwise Elo

GPT-6 Astra1,498.47 Elo
2.1

GPT-6 Astra (Medium); 220 tasks; Stirrup agent; pairwise Elo

Muse Spark 1.21,480.04 Elo
2.1

Muse Spark 1.2 (Xhigh); 220 tasks; Stirrup agent; pairwise Elo

Claude Fable 5.11,476.7 Elo
2.1

Claude Fable 5.1 (Low, Default Fallback); 220 tasks; Stirrup agent; pairwise Elo

Claude Sonnet 51,470.38 Elo
2.1

Claude Sonnet 5 (Max); 220 tasks; Stirrup agent; pairwise Elo

MiniMax M3.1 Flash1,465.36 Elo
2.1

MiniMax-M3.1-Flash; 220 tasks; Stirrup agent; pairwise Elo

Claude Opus 4.81,461.75 Elo
2.1

Claude Opus 4.8 (Max); 220 tasks; Stirrup agent; pairwise Elo

GPT-6.1 Sol1,459.05 Elo
2.1

GPT-6.1 Sol (Medium); 220 tasks; Stirrup agent; pairwise Elo

2.1

DeepSeek V4 Pro 0813 (Max); 220 tasks; Stirrup agent; pairwise Elo

Gemini 3.8 Flash1,445.43 Elo
2.1

Gemini 3.8 Flash (High); 220 tasks; Stirrup agent; pairwise Elo

GPT-5.6 Sol1,431.69 Elo
2.1

GPT-5.6 Sol (Medium); 220 tasks; Stirrup agent; pairwise Elo

Gemini 3.8 Flash1,431.32 Elo
2.1

Gemini 3.8 Flash (Medium); 220 tasks; Stirrup agent; pairwise Elo

GPT-6 Astra1,403.26 Elo
2.1

GPT-6 Astra (Low); 220 tasks; Stirrup agent; pairwise Elo

GLM-5.21,381.49 Elo
2.1

GLM-5.2 (Max); 220 tasks; Stirrup agent; pairwise Elo

Claude Sonnet 51,366.08 Elo
2.1

Claude Sonnet 5 (Xhigh); 220 tasks; Stirrup agent; pairwise Elo

Claude Opus 4.71,361.69 Elo
2.1

Claude Opus 4.7 (Max); 220 tasks; Stirrup agent; pairwise Elo

GPT-5.51,359.24 Elo
2.1

GPT-5.5 (Xhigh); 220 tasks; Stirrup agent; pairwise Elo

2.1

DeepSeek V4.1 Flash (Non-reasoning); 220 tasks; Stirrup agent; pairwise Elo

GPT-5.51,335.86 Elo
2.1

GPT-5.5 (High); 220 tasks; Stirrup agent; pairwise Elo

GLM-5.31,333.34 Elo
2.1

GLM-5.3 (Low); 220 tasks; Stirrup agent; pairwise Elo

Claude Sonnet 5.51,328.63 Elo
2.1

Claude Sonnet 5.5 (Medium, Default Fallback); 220 tasks; Stirrup agent; pairwise Elo

GPT-6.1 Sol1,321.98 Elo
2.1

GPT-6.1 Sol (Low); 220 tasks; Stirrup agent; pairwise Elo

Gemini 3.8 Flash1,316.07 Elo
2.1

Gemini 3.8 Flash (Low); 220 tasks; Stirrup agent; pairwise Elo

GPT-5.6 Sol1,312.98 Elo
2.1

GPT-5.6 Sol (Low); 220 tasks; Stirrup agent; pairwise Elo

Claude Opus 51,310.56 Elo
2.1

Claude Opus 5 (Low); 220 tasks; Stirrup agent; pairwise Elo

Claude Sonnet 51,262.64 Elo
2.1

Claude Sonnet 5 (High); 220 tasks; Stirrup agent; pairwise Elo

GLM-5.21,251.81 Elo
2.1

GLM-5.2 (Non-reasoning); 220 tasks; Stirrup agent; pairwise Elo

MiniMax M31,249.95 Elo
2.1

MiniMax-M3; 220 tasks; Stirrup agent; pairwise Elo

GPT-5.6 Sol1,249.91 Elo
2.1

GPT-5.6 Sol (Non-reasoning); 220 tasks; Stirrup agent; pairwise Elo

GPT-5.51,242.99 Elo
2.1

GPT-5.5 (Medium); 220 tasks; Stirrup agent; pairwise Elo

Claude Opus 5.51,240.15 Elo
2.1

Claude Opus 5.5 (Low, Default Fallback); 220 tasks; Stirrup agent; pairwise Elo

Claude Sonnet 51,220.8 Elo
2.1

Claude Sonnet 5 (Non-reasoning, High); 220 tasks; Stirrup agent; pairwise Elo

Muse Spark 1.11,217.25 Elo
2.1

Muse Spark 1.1 (Xhigh); 220 tasks; Stirrup agent; pairwise Elo

Claude Sonnet 5.51,181.72 Elo
2.1

Claude Sonnet 5.5 (Low, Default Fallback); 220 tasks; Stirrup agent; pairwise Elo

Claude Sonnet 51,166.62 Elo
2.1

Claude Sonnet 5 (Medium); 220 tasks; Stirrup agent; pairwise Elo

2.1

DeepSeek V4 Pro 0424 (Max); 220 tasks; Stirrup agent; pairwise Elo

2.1

DeepSeek V4 Pro 0424 (High); 220 tasks; Stirrup agent; pairwise Elo

Kimi K31,135.77 Elo
2.1

Kimi K3 (Low); 220 tasks; Stirrup agent; pairwise Elo

Claude Sonnet 51,075.82 Elo
2.1

Claude Sonnet 5 (Low); 220 tasks; Stirrup agent; pairwise Elo

2.1

DeepSeek V4 Pro 0813 (Non-reasoning); 220 tasks; Stirrup agent; pairwise Elo

GPT-5.51,053.07 Elo
2.1

GPT-5.5 (Low); 220 tasks; Stirrup agent; pairwise Elo

Kimi K2.61,046.53 Elo
2.1

Kimi K2.6 (Reasoning); 220 tasks; Stirrup agent; pairwise Elo

Muse Spark1,006.05 Elo
2.1

Muse Spark; 220 tasks; Stirrup agent; pairwise Elo

GPT-5.5977.8 Elo
2.1

GPT-5.5 (Non-reasoning); 220 tasks; Stirrup agent; pairwise Elo

Kimi K2.5846.64 Elo
2.1

Kimi K2.5 (Reasoning); 220 tasks; Stirrup agent; pairwise Elo

Gemini 3.1 Pro798.88 Elo
2.1

Gemini 3.1 Pro Preview; 220 tasks; Stirrup agent; pairwise Elo

gpt-oss-120b615.64 Elo
2.1

gpt-oss-120b (High); 220 tasks; Stirrup agent; pairwise Elo

gpt-oss-20b347.19 Elo
2.1

gpt-oss-20b (High); 220 tasks; Stirrup agent; pairwise Elo

gpt-oss-120b179.43 Elo
2.1

gpt-oss-120b (Low); 220 tasks; Stirrup agent; pairwise Elo