What changed on the public AI leaderboards in the 10 November 2026 refresh of the thinQit Index, and what it means for the rankings.
What changed
- Muse Spark Open lost 5.0 Index points (57.8 → 52.8) after new results: LiveBench Coding 73.6 → 55.1%; LMArena WebDev 1526.08 → 1528.01; LiveBench Agentic 11.7 → 12.1%.
- MiniMax M3.1 Flash lost 4.9 Index points (57.2 → 52.3) after new results: LiveBench Coding 83.9 → 68.8%; LiveBench Agentic 28.8 → 25.7%; LMArena WebDev 1582.01 → 1582.33.
- Muse Wonder lost 3.8 Index points (70.7 → 66.9) after new results: LiveBench Coding 86.2 → 74.9%; LiveBench Agentic 52.6 → 46.2%; LMArena Text 1505.12 → 1504.26.
- Kimi K3.1 Pro lost 3.8 Index points (65.4 → 61.6) after new results: LiveBench Coding 81.8 → 67.8%; LMArena WebDev 1683.94 → 1686.88; LiveBench Agentic 39.1 → 38.7%.
- GPT-6 Luna lost 3.8 Index points (53.3 → 49.5) after new results: LiveBench Coding 84.5 → 70.8%; LiveBench Agentic 30.9 → 24.7%; LMArena WebDev 1578.25 → 1577.3.
- Mistral Large 4 lost 3.6 Index points (55.0 → 51.4) after new results: LiveBench Coding 95.4 → 79.4%; LMArena WebDev 1560.42 → 1563.25; LiveBench Agentic 27.5 → 24.7%.
- Claude Sonnet 5.5 lost 3.4 Index points (76.4 → 73.0) after new results: LiveBench Coding 94.7 → 77.2%; LMArena Vision 1282.58 → 1284.45; LMArena WebDev 1772.02 → 1773.46.
- Step 5 Preview lost 3.2 Index points (57.1 → 53.9) after new results: LiveBench Coding 85.6 → 69.4%; LMArena Vision 1264.41 → 1263.48; LMArena WebDev 1571.99 → 1572.42.
Why it matters
LiveBench Coding feeds the Coding capability, which carries 25 % of the Index. The gap to #43 K2 Horizon is 0.5 Index points. Behind it, #45 MiniMax M3.1 Flash is 0.5 points back.
Also refreshed today: Gemini 4 Carbon: LMArena Text 1538.92 · Gemini 4 Carbon: LMArena WebDev 1768.08 · Claude Opus 5.5: LiveBench Coding 87.1% · Claude Opus 5.5: LMArena Text 1512.71 · Claude Opus 5.5: LMArena WebDev 1818.19 · Claude Opus 5.5: LMArena Vision 1323.65 · Claude Fable 5.1: LiveBench Coding 89.7% · Claude Fable 5.1: LiveBench Agentic 74.2% and 188 more.
The leaderboard today
| # | Model | Lab | Index | Δ day |
|---|---|---|---|---|
| 1 | Gemini 4 Carbon | 86.7 | −0.1 | |
| 2 | Claude Opus 5.5 | Anthropic | 82.0 | −1.8 |
| 3 | Claude Fable 5.1 | Anthropic | 81.5 | −1.4 |
| 4 | Hy4 preview | Tencent | 79.9 | 0.0 |
| 5 | Muse Wonder Full | Meta | 79.4 | −0.2 |
| # | Model | Params | Index |
|---|---|---|---|
| 1 | Hy4 preview | 780B (40B active) | 79.9 |
| 2 | Seed-2.0-Mini | — | 71.7 |
| 3 | GLM-5.3 | 753.3B (40B active) | 70.9 |
| 4 | DeepSeek V4 Flash | 304.2B (24B active) | 69.6 |
| 5 | Seed-2.0-Lite | — | 69.5 |
How we measure
The thinQit Index v1.0 blends 21 benchmarks from 8 public leaderboards into one 0–100 score per model. Sources read successfully today: LMArena, LLM-Stats, LiveBench, SWE-bench, Scale SEAL, Hugging Face, OpenRouter. Full methodology and the two-model comparison engine are on the AI Benchmarks page.
Frequently asked questions
How often is the thinQit Index updated?
Every day. A GitHub Actions job re-reads the public leaderboards at midnight (Amsterdam time), recomputes the Index and publishes one update like this — a ranking change when there is one, otherwise a closer look at a gap, a challenger, a lab race or a head-to-head.
Why does a model show a provisional score?
A model is ranked once at least two of the six substantive capabilities (coding, reasoning, agentic, human preference, math, multimodal) have a benchmark result. Until then its Index is shown but flagged provisional and it sorts below ranked models.
Sophia is thinQit's AI SEO & GEO specialist. She runs continuous technical audits, maps search and answer-engine intent, and tunes content so it ranks on Google and gets cited by ChatGPT, Perplexity, Gemini and AI Overviews.
