Benchmarks
357 sources · 1,263 models × 3,079 benchmarks · 162,127 scoresLiveBench 2.0 overall for the 2026-11-06 release: the weighted mean of its seven suite scores, Long Agentic x2, Coding and Design x1.5, and Reasoning, Browser, Data Analysis and Research x1, as the release's categories file sets. LiveBench 2.0's release of 2026-11-06 (Abacus AI) reruns Long Agentic, Coding, Research and Design on new tasks and keeps the Reasoning and Data Analysis runs and the October 25 Browser runs; each suite is one run per model at its highest supported effort, a task the run could not measure scores 0, and a model without an input a suite needs scores 0 there. A separate family from the 2026-10-14 release, whose Overall weights suites equally.
66th strongest of 818 fitted benchmarks
Signal-to-noise +16.23
360th hardest of 818 fitted benchmarks
Capability needed 154.8
| # | Model | Setting | Score | Source |
|---|---|---|---|---|
| 1 | Gemini 4 Neon | x-high | 87.4 | |
| 2 | Claude Fable 5.2 | x-high | 84.9 | |
| 3 | GPT-6.2 Astra | max | 84.9 | |
| 4 | GPT-6 Astra | max | 84.3 | |
| 5 | Claude Fable 5.1 | max | 84.0 | |
| 6 | Claude Opus 5.5 | max | 83.4 | |
| 7 | Gemini 4 Carbon | high | 83.3 | |
| 8 | GPT-6.1 Sol | max | 81.2 | |
| 9 | GPT-6.2 Sol | max | 80.3 | |
| 10 | GPT-6.1 Terra | max | 78.6 | |
| 11 | Gemini 4 Argon | high | 77.7 | |
| 12 | Gemini 4 Hydrogen | x-high | 77.0 | |
| 13 | Claude Sonnet 5.5 | max | 76.8 | |
| 14 | Qwen4-Max | x-high | 73.1 | |
| 15 | Muse Wonder | max | 70.7 | |
| 16 | DeepSeek V4.1 Flash 1017 | max | 70.3 | |
| 17 | Claude Haiku 5.5 | max | 68.9 | |
| 18 | Kimi K3.1 Pro | max | 68.7 | |
| 19 | Muse Spark 1.3 | max | 68.6 | |
| 20 | Grok 4.7 | x-high | 67.9 | |
| 21 | GPT-5.6 Terra | max | 67.2 | |
| 22 | MiMo-V2.6-Pro | high | 66.1 | |
| 23 | Kimi K3 | max | 65.0 | |
| 24 | MiniMax M3.1 Flash | max | 63.7 | |
| 25 | Grok 4.6 | x-high | 63.5 | |
| 26 | GLM-5.3 Flash | max | 60.7 | |
| 27 | GPT-6 Luna | max | 60.4 | |
| 28 | Mistral Large 4 | high | 60.3 | |
| 29 | Gemini 3.8 Flash | high | 57.3 | |
| 30 | Step 5 Preview | high | 54.9 | |
| 31 | Qwen3.8 27B | x-high | 54.0 | |
| 32 | Muse Spark Open | x-high | 48.5 | |
| 33 | Nemotron 3.5 Super VL 120B A12B BF16 | thinking | 47.0 | |
| 34 | Inkling | not reported | 44.3 | |
| 35 | Nemotron 3.5 Super VL 120B A12B NVFP4 | thinking | 42.3 | |
| 36 | Gemma 4 31B IT | max | 32.4 |
Benchmarks outside its family whose model rankings track it most reliably (Spearman ρ over at least 10 shared models).