The independent leaderboard

Benchmarks, weighed honestly

Every lab’s own slides put its newest model on top. Independent evaluators run the same tests on everyone, publish their method, and can’t quietly cherry-pick — so when the scoring is done by someone with nothing to sell, the ranking changes. These charts redraw that third-party record in the Omni Labs hand.

The headline index

Which models are actually smartest

The Artificial Analysis Intelligence Index — one composite of roughly ten independent evals — for every model that reports it, ranked highest first.

Independent data · Artificial Analysis
The color key · by lab family
Claude · Mistral · MiniMaxGeminiGPT · Llama · DeepSeekGLMGrok · Qwen · Kimi

With more labs than accent colors, some families share a hue — color is a coarse family key, always read beside the name.

Quality versus price

Where the value actually is

Intelligence against blended price per million tokens (a 3-to-1 input/output blend, log scale, spanning ~150×). Up and to the left is the sweet spot; the dashed line traces the efficiency frontier — nothing offers more for less.

Independent data · Artificial Analysis
Benchmarks versus reality

The fall from Verified to Pro

A model’s SWE-bench Verified score is the leaderboard number. SWE-bench Pro — harder and contamination-resistant — is closer to the job. The drop between them, sorted largest-first, shows whose coding scores don’t generalise.

Independent data · SWE-bench Verified & SWE-bench Pro (Scale AI)
SWE-bench VerifiedSWE-bench Prothe drop → reality
Agentic reliability

Getting real work done in a terminal

Terminal-Bench Hard runs long-horizon agentic tasks end-to-end in a real shell — editing files, running tools, recovering from errors. It is the closest thing to the job, and even the frontier sits far below saturation. Independently scored, ranked highest first.

Independent data · Artificial Analysis (Terminal-Bench Hard)
Coding, contamination-resistant

The benchmark that closed the loophole

Datacurve built DeepSWE — 113 original tasks across 91 live repos — after finding SWE-bench Pro shipped full .git history in its containers, letting some models read the gold patch straight from the log (Opus 4.6/4.7 did, on a double-digit share of passes). DeepSWE ships shallow clones, verifies with a 0.3% false-positive rate, and spreads the field back out — the frontier sits near 70%, nowhere close to saturated.

Independent data · Datacurve (DeepSWE)

These are our own visualizations of publicly reported, independent third-party data — redrawn in our house style, not reproductions of any provider’s or aggregator’s charts. Figures reflect the dataset at time of writing; follow the live boards above for the current record.