Artificial Analysis Agentic Index (AA Agentic Index)
We show this table for reference; we do not rank on it.
A display-only Artificial Analysis agentic index.
Benchmark score on AA Agentic Index — October 29, 2026
We compile the AA Agentic Index rows from secondary reports. Claude Fable 5.1 leads the table at 58.0%, followed by Claude Opus 5 (56.2%) and Muse Spark 1.3 (55.7%). We do not use these results to rank models overall.
Claude Fable 5.1
Anthropic
Claude Opus 5
Anthropic
Muse Spark 1.3
Meta
79 modelsAgenticCurrentDisplay onlyUpdated October 29, 2026
Benchmark score table (79 models)
| Rank | Model / configuration | Score | Parameters (B) | Open / closed |
|---|---|---|---|---|
| 1 | Claude Fable 5.1Anthropic | 58.0% | Not reported | Closed |
| 2 | Claude Opus 5Anthropic | 56.2% | Not reported | Closed |
| 3 | Muse Spark 1.3Meta | 55.7% | Not reported | Closed |
| 4 | GLM-5.3Z.AI | 53.4% | Not reported | Open |
| 5 | Grok 4.6xAI | 53.4% | Not reported | Closed |
| 6 | GPT-6 AstraOpenAI | 51.5% | Not reported | Closed |
| 7 | Claude Fable 5Anthropic | 51.0% | Not reported | Closed |
| 8 | Kimi K3Moonshot AI | 50.6% | Not reported | Open |
| 9 | GPT-5.6 SolOpenAI | 50.5% | Not reported | Closed |
| 10 | Qwen3.8 Max PreviewAlibaba | 49.6% | Not reported | Closed |
| 11 | DeepSeek V4 Pro 0813DeepSeek | 49.6% | Not reported | Open |
| 12 | Qwen3.8-27BAlibaba | 46.5% | Not reported | Open |
| 13 | Claude Sonnet 5Anthropic | 44.3% | Not reported | Closed |
| 14 | Muse Spark 1.2Meta | 44.0% | Not reported | Closed |
| 15 | GPT-5.6 TerraOpenAI | 43.7% | Not reported | Closed |
| 16 | GPT-5.6 LunaOpenAI | 42.7% | Not reported | Closed |
| 17 | Claude Opus 4.8Anthropic | 42.6% | Not reported | Closed |
| 18 | Grok 4.5xAI | 42.1% | Not reported | Closed |
| 19 | DeepSeek V4 Flash 0731DeepSeek | 41.7% | Not reported | Open |
| 20 | Gemini 3.8 FlashGoogle | 41.1% | Not reported | Closed |
| 21 | Claude Opus 4.7 (Adaptive)Anthropic | 39.5% | Not reported | Closed |
| 22 | GLM-5.2Z.AI | 39.4% | Not reported | Open |
| 23 | GPT-5.5OpenAI | 37.3% | Not reported | Closed |
| 24 | Gemini 3.7 FlashGoogle | 36.4% | Not reported | Closed |
| 25 | DeepSeek V4 Pro (High)DeepSeek | 35.3% | Not reported | Open |
| 26 | Quasar 438BMultiverse Computing | 32.7% | Not reported | Closed |
| 27 | MiniMax M3MiniMax | 30.8% | Not reported | Open |
| 28 | Gemini 3.6 FlashGoogle | 30.1% | Not reported | Closed |
| 29 | Muse Spark 1.1Meta | 27.5% | Not reported | Closed |
| 30 | Gemini 3.5 FlashGoogle | 27.3% | Not reported | Closed |
| 31 | Hy3Tencent | 25.6% | Not reported | Open |
| 32 | Hy3 PreviewTencent | 25.6% | Not reported | Open |
| 33 | GLM-5.1Z.AI | 25.2% | Not reported | Open |
| 34 | Inkling-SmallThinking Machines Lab | 24.9% | Not reported | Open |
| 35 | InklingThinking Machines Lab | 24.3% | Not reported | Open |
| 36 | Qwen3.7 MaxAlibaba | 23.9% | Not reported | Closed |
| 37 | MiMo-V2.5-ProXiaomi | 22.7% | Not reported | Closed |
| 38 | Kimi K2.7 CodeMoonshot AI | 22.5% | Not reported | Open |
| 39 | Kimi K2.6Moonshot AI | 22.1% | Not reported | Open |
| 40 | Nemotron 3 UltraNVIDIA | 21.7% | Not reported | Open |
| 41 | Ling 3.0 FlashInclusionAI | 21.0% | Not reported | Open |
| 42 | Qwen3.6-27BAlibaba | 20.1% | Not reported | Open |
| 43 | Qwen3.7 PlusAlibaba | 19.7% | Not reported | Closed |
| 44 | GPT-5.4 miniOpenAI | 19.6% | Not reported | Closed |
| 45 | GPT-5.4 nanoOpenAI | 17.7% | Not reported | Closed |
| 46 | Grok 4.3xAI | 17.2% | Not reported | Closed |
| 47 | MiniMax M2.7MiniMax | 16.8% | Not reported | Open |
| 48 | Gemini 3.5 Flash-LiteGoogle | 15.9% | Not reported | Closed |
| 49 | Qwen3.6-35B-A3BAlibaba | 15.0% | Not reported | Open |
| 50 | Muse Glimmer 30BMeta | 10.5% | Not reported | Open |
| 51 | Gemini 3.1 ProGoogle | 10.3% | Not reported | Closed |
| 52 | Qwen3.5-122B-A10BAlibaba | 9.6% | Not reported | Open |
| 53 | Mistral Medium 3.5 128BMistral | 9.3% | Not reported | Open |
| 54 | GPT-5 miniOpenAI | 8.9% | Not reported | Closed |
| 55 | Gemma 4 31BGoogle | 6.7% | Not reported | Open |
| 56 | GPT-OSS 120BOpenAI | 6.2% | Not reported | Open |
| 57 | Nemotron 3.5 Lightning 30B A3B NVFP4NVIDIA | 6.1% | Not reported | Open |
| 58 | Mercury 2Inception | 4.0% | Not reported | Closed |
| 59 | Granite 4.2 8BIBM | 3.7% | Not reported | Open |
| 60 | Command A+Cohere | 3.6% | Not reported | Open |
| 61 | Gemini 2.5 ProGoogle | 3.5% | Not reported | Closed |
| 62 | Mistral Large 3Mistral | 2.4% | Not reported | Closed |
| 63 | Mistral Small 4Mistral | 1.4% | Not reported | Open |
| 64 | Mistral Small 4 (Reasoning)Mistral | 1.4% | Not reported | Open |
| 65 | GPT-OSS 20BOpenAI | 1.4% | Not reported | Open |
| 66 | Trinity-Large-ThinkingArcee AI | 1.2% | Not reported | Open |
| 67 | Trinity-Large-PreviewArcee AI | 1.2% | Not reported | Open |
| 68 | Ministral 3 14B (Reasoning)Mistral | 1.1% | Not reported | Open |
| 69 | Ministral 3 14BMistral | 1.1% | Not reported | Open |
| 70 | Nemotron 3 Nano 30BNVIDIA | 1.0% | Not reported | Open |
| 71 | DeepSeek V3DeepSeek | 0.8% | Not reported | Open |
| 72 | Ministral 3 3B (Reasoning)Mistral | 0.8% | Not reported | Open |
| 73 | Ministral 3 3BMistral | 0.8% | Not reported | Open |
| 74 | Celeris-1Celeris | 0.7% | Not reported | Closed |
| 75 | Llama 4 MaverickMeta | 0.6% | Not reported | Open |
| 76 | Ministral 3 8B (Reasoning)Mistral | 0.6% | Not reported | Open |
| 77 | Ministral 3 8BMistral | 0.6% | Not reported | Open |
| 78 | Llama 4 ScoutMeta | 0.6% | Not reported | Open |
| 79 | Gemma 3 27BGoogle | 0.1% | Not reported | Open |
Among the reported AA Agentic Index rows, Claude Fable 5.1 is first at 58.0%. The third row is 2.3 points behind. The broader top-10 range is 8.4 points, so many of the published results sit in a relatively narrow band.
79 models have been evaluated on AA Agentic Index. The benchmark falls in the Agentic category. AA Agentic Index is currently displayed for reference but excluded from the scoring formula, so it does not directly affect overall rankings.
About AA Agentic Index
Year
2026
Tasks
Cross-benchmark agentic index
Format
Aggregated model score
Difficulty
Display-only external reference
BenchLM mirrors this agentic index for comparison, but does not use it as a weighted agentic benchmark row.
Freshness and provenance
Version
AA Agentic Index 2026
Refresh cadence
Quarterly
Staleness state
Current
Question availability
Public benchmark set
BenchLM uses freshness metadata to decide whether a benchmark should still be treated as a strong differentiator, a benchmark to watch, or a display-only reference. For the full scoring policy, see the BenchLM methodology page.
Questions
What does AA Agentic Index measure?
A display-only Artificial Analysis agentic index.
Which model scores highest on AA Agentic Index?
Claude Fable 5.1 by Anthropic currently leads with a score of 58.0% on AA Agentic Index.
How many models are evaluated on AA Agentic Index?
79 AI models have published results on AA Agentic Index in the BenchLM catalog.
Compare top models on AA Agentic Index
Know when it’s worth switching models
The model to choose, the cheaper alternative, and the release we would wait on.
Read a sample issueJoin 6,100+ subscribers.
One email each week. Unsubscribe anytime.