About
llminfo IQ
llminfo IQ combines benchmark results to compare models of the same type. A score of 100 is average for that group.
Each 15 points is one standard deviation. Current frontier models score around 150–175. Chat models are compared with chat models, and embedding models with embedding models. The scores draw on ~90,000 benchmark results, and each one shows how much data sits behind it.
Each benchmark is weighted by how much signal it actually carries. The method is open. We publish the full per-benchmark weight table with every refresh. Every refresh, we also check the index against independent aggregate ratings. The main check is the Epoch Capabilities Index (Spearman ρ ≈ 0.98). The second is the Artificial Analysis Intelligence Index, as relayed by third-party sources. Any quoted agreement with an Artificial Analysis figure names the source that relayed it, never Artificial Analysis as our own source.
Last calculated 22 Nov 2026. A model's score can change over time as new benchmark results arrive and the pool of models shifts. It's always shown with its date.
What it is not
- Not a measurement of any single benchmark. It combines results from many tests. A high-IQ model can still lose a particular benchmark to a lower-IQ one.
- Not comparable across model types. Chat models are ranked among chat models, embedding models among embedding models. An embedding IQ of 130 and a chat IQ of 130 are unrelated claims.
- Not a quality guarantee for a domain. A single number can't capture that a model is unusually good at coding for its general ability. Coding has its own index, Coding IQ, fitted from coding benchmarks on a separate scale. The two numbers aren't comparable, and their difference doesn't measure specialisation.
- Not immune to gaps in the data. A score with little data behind it carries real uncertainty. Models tested only on niche community boards are pulled toward the average on purpose.
- Not human IQ. The name borrows the scale convention (100 ± 15) for readability, nothing else.
How much data sits behind a score
Select a score in the table to see how much data sits behind it. A score with little data behind it is a rough estimate and is shown as one.
- highPlenty of data behind it: 10 or more benchmarks, including at least 2 from trusted-tier sources.
- mediumA fair amount of data behind it: 5 or more benchmarks.
- lowLittle data behind it: 3 or 4 benchmarks, shown as a rough estimate (≈ N, no decimals).
Current weight table
The benchmarks that drive the chat score most right now, by measured weight. A benchmark’s weight is how well it separates strong models from weak ones. Weights are re-measured from the data on every refresh. Nothing is hand-picked, so this table moves.
| Benchmark | Weight | Models |
|---|---|---|
| Vals SRE Bench | 4.34 | 34 |
| Vals Terminal-Bench-Science | 4.18 | 46 |
| Blueprint Bench 2 Raw score | 4.03 | 40 |
| SWE-bench Verified (Epoch) | 3.61 | 31 |
| Vals SWE-bench Multimodal | 3.60 | 21 |
| SWE-Bench Pro | 3.48 | 92 |
| Blueprint Bench 2 Score | 3.46 | 40 |
| OpenHands Index: swe-bench | 3.46 | 31 |
| Vals ProgramBench | 3.42 | 64 |
| NYT Connections – Older models (Lech Mazur) | 3.37 | 77 |
| VideoMMMU | 3.33 | 27 |
| Vals SWE-bench | 3.33 | 83 |
| LiveCodeBench Easy | 3.32 | 16 |
| Deep20Bench success rate | 3.27 | 28 |
| Multi-SWE-bench JavaScript Verified | 3.27 | 11 |
| MathVista-Mini | 3.24 | 23 |
| Vals EMB | 3.14 | 73 |
| Multi-SWE-Bench | 3.12 | 10 |
| Vals AIME | 3.09 | 81 |
| GDPval | 3.09 | 10 |
| Arena Agent Tool Hallucination | 3.08 | 62 |
| Vals Router Index | 3.08 | 24 |
| Humanity's Last Exam (no tools, text-only) | 3.04 | 15 |
| FrontierMath Tier 4 | 3.01 | 47 |
| GSO | 2.99 | 30 |
| METR Time Horizons | 2.99 | 35 |
| SWE-bench Multilingual | 2.97 | 58 |
| FrontierCode 1.1 | 2.97 | 41 |
| GDM-MRCRv2: 8 needle 8k | 2.96 | 68 |
| Vals Index | 2.93 | 40 |
| BrowserGym WorkArena-L2 | 2.92 | 11 |
| Vals EMB v1.1 | 2.89 | 44 |
| Vals LiveCodeBench | 2.89 | 127 |
| LiveCodeBench Medium | 2.89 | 16 |
| INCLUDE-base-44 0-shot greek (Speakleash) | 2.89 | 15 |
| BFCL-v3 | 2.88 | 16 |
| LiveBench theory_of_mind | 2.85 | 51 |
| LiveBench zebra_puzzle | 2.84 | 51 |
| VideoMME w sub. | 2.84 | 10 |
| PutnamBench lean-wsolution solved | 2.83 | 12 |
Other model types (embedding, image generation, video generation, text to speech)
embedding
| Benchmark | Weight | Models |
|---|---|---|
| MTEB(fas, v2) - Clustering | 1.88 | 12 |
| ViDoRe(v3) - Public split | 1.82 | 6 |
| ViDoRe(v3) | 1.80 | 6 |
| MTEB(fra, v1) | 1.72 | 17 |
| MTEB(fra, v1) - Clustering | 1.65 | 22 |
| MTEB(fra, v1) - Mean (TaskType) | 1.65 | 17 |
| BEIR-NL | 1.63 | 5 |
| MTEB(kor, v1) | 1.63 | 21 |
image generation
| Benchmark | Weight | Models |
|---|---|---|
| UniGenBench++ (Chinese Long) - Color | 2.51 | 25 |
| UniGenBench++ (Chinese) - Animal | 2.44 | 28 |
| UniGenBench++ (Chinese Long) - Layout-Overall | 2.40 | 25 |
| UniGenBench++ (Chinese) - Overall | 2.37 | 28 |
| UniGenBench++ (Chinese Long) - Material | 2.33 | 25 |
| UniGenBench++ (Chinese Long) - Attribute-Overall | 2.31 | 25 |
| UniGenBench++ (Chinese) - Feature matching | 2.28 | 28 |
| UniGenBench++ (Chinese) - 2D | 2.27 | 28 |
video generation
| Benchmark | Weight | Models |
|---|---|---|
| HumanScore | 2.18 | 9 |
| HumanScore Kinematic | 2.10 | 9 |
| Design Arena Elo (models/multitovideo) | 1.89 | 14 |
| Physics-IQ Verified | 1.73 | 10 |
| Arena ELO Text to Video | 1.38 | 39 |
| llm-stats Video Generation Arena | 1.38 | 24 |
| Arena ELO Video Edit | 1.36 | 10 |
| Design Arena Elo (models/imagetovideo) | 1.33 | 26 |
text to speech
| Benchmark | Weight | Models |
|---|---|---|
| Design Arena Elo (models/audiorealism) | 1.32 | 10 |
| Pashto ASR FLEURS ps_af CER (zero-shot) | 1.24 | 7 |
| Vals VoiceCodeBench | 1.23 | 10 |
| MAEB(v1, audio-only) - Mean (TaskType) | 1.23 | 5 |
| EmergentTTS-Eval Syntactic Complexity WER | 1.16 | 7 |
| EmergentTTS-Eval Questions WER | 1.15 | 7 |
| Pashto ASR FLEURS ps_af WER (zero-shot) | 1.11 | 7 |
| Pashto ASR CV24_filtered WER (zero-shot) | 1.09 | 7 |