Moonshot AIChina· released Nov 2025· open weights
Provisional: 10 comparable tests so far. Ranked once it reaches 6 across 3 counting categories.
Moonshot's open-weight K2 line; the Thinking release is the variant publishers test.
Snapshot September 8, 2026
OpenCharts Index
60.6
Provisional: not enough comparable tests or counting categories to rank yet · 10 comparable tests · 0 counting categories
Drop any single test and the Index lands between 55.9 and 64.6.
Needs more results (2 of 8 so far)
Averaging 49.8 on the 2 it has; the score opens at 4 of 8.
Distance to today's frontier on the hardest tests. Not a measure of intelligence.
What it's known for
Its widest gap is on APEX-Agents, 91.4% behind the best published result. No comparable results yet in knowledge and multimodal. Between 35% and 40% behind the best published results, on average.
Strongest category
No category counts yet: a category counts once it has more than one comparable test and the model has taken at least half of them. The scores it has are below.
Best single result
Answer a PhD-level multiple-choice question in physics, chemistry or biology that a web search will not settle.
84.2%
83.7% of the best (95.8%), above chance · #57 of 78.
Expert-level science questions are the closest thing to asking a specialist colleague. A model that gets them right can check a technical claim instead of echoing it.
Where it trails
Complete professional-services work in consulting, law and finance as an agent, graded against expert rubrics.
4.1%
91.4% behind the best published result (47.4%) · #47 of 49.
Expert-graded knowledge work is what professional teams would actually delegate.
Not measured yet
No comparable result yet in these 2 categories. A blank is a blank, never a zero, and it does not lower the Index.
Covered, but not averaged into the Index:
See it at work
Every test it has taken, by category, with the kind of task it faced, who produced the number and where the result landed against the best published one. The examples are original and illustrative, never items from the datasets themselves.
Reasoning · mean 83.7 across 1 comparable test · not counted toward the Index yet (fewer than half the category's tests)
Answer a PhD-level multiple-choice question in physics, chemistry or biology that a web search will not settle.
84.2%
83.7% of the best (95.8%), above chance
#57 of 78 · n = 198
Frontier models cluster in the eighties and nineties; Epoch's standard errors on this test are 2 to 3 points.
Source: Epoch AI (internal runs)Every result is read against the best published one on its test: percent scores as a share of the best above chance, Arena ratings as the expected win rate against the board leader, open-ended values on a log scale. Sitting a harder exam never lowers a model. Every number links to the publisher that produced it.
Head to head
Choose a rival. Every counted test both have taken appears side by side as a share of the best published result on that test, with the category means above. A win is a higher share; a gap under 2 points is a tie, because one item on a 45-problem exam is 2.2 points and publishers' own standard errors are of that order.
Kimi K2
provisional·Index 60.6C
0 : 5
wins · 0 ties · 5 shared
GPT-6 Astra
#1·Index 99.0A+
GPT-6 Astra comes out ahead on 5 of the 5 tests both have taken.
Category means
Reasoning
Knowledge
Coding
Math
Agentic
Human preference
Long context
Shared tests · biggest gaps first
Bars are distances to the best published result on each test (percent scores above chance, Arena ratings as win rate against the leader, open-ended values on a log scale). Raw values as the publishers report them.
Open GPT-6 AstraScores
The best published run per benchmark, the raw value as the publisher reports it, who produced the number, how close it comes to the best published result on that test, and where the model sits among every model scored on it. Percent scores are read above chance, Arena ratings as a win rate against the board leader, open-ended values on a log scale. A test with too few models to compare against is shown but not counted.
Category means
Mean · tests taken of the category's comparable tests. A faded category either rests on fewer than half of them or is a single-test category (marked “beside”), which is shown next to the Index and never averaged into it.
84.2%
83.7% of the best (95.8%), above chance
27th percentile of 78 · n = 198
59.1%
67.2% of the best (88.0%)
25th percentile of 5 · n = 225
42.8%
46.1% of the best (92.9%)
15th percentile of 69
thinking
Source: Epoch AI — AI Benchmarking Hub1,322
6% expected win rate against the board leader (1,797)
13th percentile of 86
board 2026-09-05
Source: Arena (LMArena) — Leaderboard Dataset83.1%
83.1% of the best (100.0%)
19th percentile of 78 · n = 45
35.7%
42.1% of the best (84.7%)
33th percentile of 40
thinking · agent: Terminus 2
Source: Epoch AI — AI Benchmarking Hub54 min
57.4% of the best (17.4 h) on a log scale
8th percentile of 14
thinking
Source: Epoch AI — AI Benchmarking Hub4.1%
8.6% of the best (47.4%)
4th percentile of 49
thinking
Source: Epoch AI — AI Benchmarking Hub1,430
39% expected win rate against the board leader (1,507)
26th percentile of 92
board 2026-09-02
Source: Arena (LMArena) — Leaderboard Dataset40.6%
40.6% of the best (100.0%)
17th percentile of 7
Top of the leaderboard
28 of the models on this page run inside Theo today. Theo picks the right one for each step and always shows which engine answered.