Flowerclaw

Intelligence leaderboard

Two boards, two kinds of "smart"

The Intelligence Index (Artificial Analysis) measures combined reasoning, coding and math performance; LMArena measures preference from real blind user votes. One is an exam, the other word-of-mouth — read both and no single narrative runs away with you.

Last updated

Intelligence Index (0–100): Artificial Analysis's weighted performance across reasoning/coding/math tasks. Top 30 cited only.

Artificial AnalysisVersion 2026-11-07License Proprietary — limited factual citation (intelligence index totals + top-30 entries only) per AA Terms of Use; no full-table reproduction, no chart/copy reuse. Re-review quarterly; if AA requires licensing, buy Commercial access or drop this field.Artificial Analysis

LMArenaVersion 2026-11-06License CC-BY-4.0LMArenaMay require overseas network

A few terms first

How to read this board

A high Intelligence Index means strong test-taking, but "good at tests" isn't "good to use" — speed, price and context length aren't in this table. When a new model storms the board, check whether it climbed via longer reasoning effort or a genuine upgrade.

LMArena scores come from real blind conversations and better reflect "is it pleasant to talk to", but Chinese coverage is thinner (the Chinese board has far fewer votes than the English main board). For Chinese writing help, read it together with the Chinese board and our writing page.

Open-weights models (flagged "open weights") can be self-hosted and fine-tuned — prefer them for sensitive data. Closed models usually have the higher ceiling. The gap between the two leaders is exactly your "self-host vs subscribe" trade-off.

→ For Chinese writing picks, see the writing page

FAQ

Why do the Intelligence Index and LMArena ranks differ?

They measure different things. The index is a weighted composite on standardized tasks (reasoning, coding, math); LMArena is human preference from blind votes — voters reward style as much as correctness. A model sitting a few places apart on the two boards is normal: it's either "capable but awkward" or "pleasant but loose". Pick by your scenario.

What does "open weights" mean — is it the same as open source?

Not quite. "Open weights" means the model parameters are downloadable, but the license may restrict commercial use or derivatives (most model licenses don't meet the OSI definition). Practically: you can self-host, fine-tune and control data flow — but read the license line by line.

Why does one model appear multiple times (High/Max…)?

Many models ship with "reasoning effort" levels: same weights, longer thinking budget — stronger, slower, pricier. Boards record each level separately, which is why Gemini 4 Carbon (High) and (Medium) both appear. Pick the tier your latency and budget allow.

How often does the data update?

LMArena syncs weekly from the official CC-BY-4.0 dataset; the Intelligence Index is manually verified and usually moves with major releases. The "last updated" badge at the top of each page and the version date next to each source are the snapshot you're looking at.

About the data

This column cites rather than reproduces: we quote only key scores and findings, each labeled with its source, version date and a link to the original — never a full-table copy of anyone's leaderboard.

Automated sources (LMArena / SWE-bench / LiveBench) sync weekly; the Artificial Analysis intelligence index and the pricing table are manually verified, with check dates noted in the changelog.

Benchmarks are not your experience. Each board measures something different and scores drift between releases — test against your own task before committing.

  • Artificial AnalysisVersion 2026-11-07License Proprietary — limited factual citation (intelligence index totals + top-30 entries only) per AA Terms of Use; no full-table reproduction, no chart/copy reuse. Re-review quarterly; if AA requires licensing, buy Commercial access or drop this field.Artificial Analysis

  • LMArenaVersion 2026-11-06License CC-BY-4.0LMArenaMay require overseas network

Written by the Flowerclaw editorial team · Reviewed by the Flowerclaw editorial team