Benchmarks
357 sources · 1,263 models × 3,079 benchmarks · 162,127 scoresLiveBench 2.0 (2026-11-06 release) Coding suite: Fix a real bug in a large production codebase frozen at a fixed snapshot; hidden tests grade the fix and a fix that breaks existing behaviour scores 0; 24 tasks. 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 for the LiveBench 2.0 (2026-11-06 release) family
Another benchmark represents this family in the fit.
| # | Model | Setting | Score | Source |
|---|---|---|---|---|
| 1 | GPT-6.2 Astra | max | 91.9 | |
| 2 | Gemini 4 Carbon | high | 91.0 | |
| 3 | Claude Fable 5.1 | max | 89.7 | |
| 4 | Gemini 4 Neon | x-high | 89.7 | |
| 5 | GPT-6.1 Terra | max | 89.2 | |
| 6 | GPT-6 Astra | max | 87.5 | |
| 7 | Claude Opus 5.5 | max | 87.1 | |
| 8 | Gemini 4 Argon | high | 85.7 | |
| 9 | Muse Spark 1.3 | max | 84.4 | |
| 10 | GPT-5.6 Terra | max | 84.2 | |
| 11 | GPT-6.2 Sol | max | 83.7 | |
| 12 | GPT-6.1 Sol | max | 82.1 | |
| 13 | DeepSeek V4.1 Flash 1017 | max | 80.4 | |
| 14 | Mistral Large 4 | high | 79.4 | |
| 15 | Claude Fable 5.2 | x-high | 79.2 | |
| 16 | Gemini 4 Hydrogen | x-high | 78.5 | |
| 17 | Claude Haiku 5.5 | max | 77.6 | |
| 18 | Claude Sonnet 5.5 | max | 77.2 | |
| 19 | Grok 4.6 | x-high | 76.7 | |
| 20 | Muse Wonder | max | 74.9 | |
| 21 | Gemini 3.8 Flash | high | 73.3 | |
| 22 | Grok 4.7 | x-high | 73.1 | |
| 23 | GLM-5.3 Flash | max | 70.9 | |
| 24 | GPT-6 Luna | max | 70.8 | |
| 25 | Qwen4-Max | x-high | 69.8 | |
| 26 | Qwen3.8 27B | x-high | 69.3 | |
| 27 | MiniMax M3.1 Flash | max | 68.8 | |
| 28 | Kimi K3.1 Pro | max | 67.8 | |
| 29 | Step 5 Preview | high | 67.6 | |
| 30 | MiMo-V2.6-Pro | high | 64.5 | |
| 31 | Muse Spark Open | x-high | 55.1 | |
| 32 | Nemotron 3.5 Super VL 120B A12B BF16 | thinking | 54.1 | |
| 33 | Kimi K3 | max | 53.1 | |
| 34 | Inkling | not reported | 48.7 | |
| 35 | Nemotron 3.5 Super VL 120B A12B NVFP4 | thinking | 34.8 | |
| 36 | Gemma 4 31B IT | max | 20.3 |
Benchmarks outside its family whose model rankings track it most reliably (Spearman ρ over at least 10 shared models).