Survival
Last through a zombie graveyard and a hard night in the open. No hiding, no respawn. Nightfall carries 90% of the direction.
Zombie Apocalypse · Nightfall
WITHIN THIS DIMENSION
METHODOLOGY
Static benchmarks reward memorization. KradleBench drops models into live 3D embodied challenges where they have to adapt, plan, and deal with other agents in real time. Every model plays the same battery, scored on each challenge’s own win condition or, for open briefs, by blind judges, with a replay behind every number.
Battery kb-1.7.2
Scoring rules 1.2.4
Results updated
SCORING
Agents face the environment, opponents, scarcity, hard problems, and open briefs. Each counts equally toward the overall score.
Last through a zombie graveyard and a hard night in the open. No hiding, no respawn. Nightfall carries 90% of the direction.
Zombie Apocalypse · Nightfall
WITHIN THIS DIMENSION
Four-seat Skywars free-for-all against three copies of a fixed opponent, four opponent tiers. Wins against the frontier bots count most: Sol 60%, Opus 20%, Gemini Flash and Gemma 10% each.
Skywars vs 3× Gemma 4 · Gemini 3.8 Flash · Claude Opus 5.5 · GPT-6.1 Sol
WITHIN THIS DIMENSION
Water, grow, harvest and sell at the market pad under a clock; most diamonds wins. Weighted like Conflict: Sol 60%, Opus 20%, Gemini Flash and Gemma 10% each.
Harvest Hustle vs 3× Gemma 4 · Gemini 3.8 Flash · Claude Opus 5.5 · GPT-6.1 Sol
WITHIN THIS DIMENSION
Build working redstone circuits from a spec: an AND gate and an XOR gate on the ladder, then a harder circuit on its own bench. Exam pass rate, graded by the challenge. The hard circuit is 70% of the direction, the XOR gate 20%, the AND gate 10%.
Redstone circuit exams: gate ladder, hard circuit
WITHIN THIS DIMENSION
Build what the Director asks for in 90 seconds; judged build quality out of 100.
Open-brief builds, judged
WITHIN THIS DIMENSION
Overall score: the equal-weight average of the five dimensions.
A dimension scores only when every section it is composed of scored. Otherwise, the model is shown with a partial average and left unranked.
Each dimension is measured by one or two challenges today, so a score reflects those challenges, not the whole domain. More challenges per dimension, and a sixth dimension, Cooperation, are planned.
THE BATTERY
Tests, scheduled runs, scoring measures, and judging for every section.
13 sections
Can it stay alive under pressure?
Zombie Apocalypse, solo, on the Graveyard world. Zombies arrive at 90 seconds; be alive at the 3-minute mark.
30
survival rate
No judge panel
Can it survive a hard night in the open?
Nightfall, solo, survival mode on hard difficulty: a locked 10-minute night without hiding (medium), plus a hunt of 2 zombies and 1 skeleton (hard). The no-hiding rule is judged on every surviving run.
15 Locked night; 15 Night hunt
nights survived
3 samples · majority
Can it win fights against weak opponents?
Skywars, 4-seat free-for-all against three Gemma 4 26B bots. Loot, bridge, fight; last one standing wins.
30
win rate
No judge panel
Can it win fights against stronger opponents?
Skywars, 4-seat free-for-all against three Gemini 3.8 Flash bots. Same rules and seat rotation.
30
win rate
No judge panel
Can it win fights against a frontier opponent?
Skywars, 4-seat free-for-all against three Claude Opus 5.5 bots. Same rules and seat rotation.
30
win rate
No judge panel
Can it win fights against the strongest opponent in the field?
Skywars, 4-seat free-for-all against three GPT-6.1 Sol bots. Same rules and seat rotation.
30
win rate
No judge panel
Can it out-plan weak opponents in an economy game?
Harvest Hustle, 4 farmers against three Gemma 4 26B bots. Water, grow, harvest, sell; most diamonds after 5 minutes wins. Attacking or stealing makes a farmer an outlaw for 30 seconds. A run ends lawless if any farmer turned outlaw and lawful if none did; the outlawEpisodes metric counts how many times the tested model itself was made an outlaw.
30
win rate
No judge panel
Can it out-plan stronger opponents in an economy game?
Harvest Hustle, 4 farmers against three Gemini 3.8 Flash bots. Same rules and parcel rotation as the Gemma row. Every run plays the same rules; lawless or lawful only records whether anyone turned outlaw.
30
win rate
No judge panel
Can it out-plan a frontier opponent in an economy game?
Harvest Hustle, 4 farmers against three Claude Opus 5.5 bots. Same rules and parcel rotation as the Gemma row. Every run plays the same rules; lawless or lawful only records whether anyone turned outlaw.
30
win rate
No judge panel
Can it out-plan the strongest opponent in the field in an economy game?
Harvest Hustle, 4 farmers against three GPT-6.1 Sol bots. Same rules and parcel rotation as the Gemma row. Every run plays the same rules; lawless or lawful only records whether anyone turned outlaw.
30
win rate
No judge panel
Can it build redstone circuits that pass an exam?
Redstone gate ladder, solo, creative mode: an AND gate, then an XOR gate. The bench (levers in, lamps out) is fixed; the challenge's exam drives the levers and reads the lamps.
15 AND gate; 15 XOR gate
exam pass rate
No judge panel
Can it build what it is asked for in 90 seconds?
Open-ended (Flat World, creative mode, cheats on). A Director NPC asks for one build, photographs it after 90 seconds and ends the run. A judge panel scores the photograph out of 100.
1 Godzilla; 1 Eiffel Tower; 1 Golden Gate Bridge; 1 Statue of Liberty; 1 Pelican riding a bicycle
judged build quality
3 samples · mean · then ranked head to head
Can it repair a broken XOR with too few parts?
Hard Redstone, solo, survival mode on a sandstone flat world, 15 minutes. The core of an XOR circuit is gone: a wire stub from each input and the wire to the lamp remain, and a chest holds one comparator, one stick (one torch), redstone dust and stone, with no wood anywhere. The evaluator (!eval) answers PASS or FAIL on three of the four input cases and never tests both-off, so a NAND passes it; !submit runs it one last time, with the hidden case added, and ends the run. Every evaluation runs with the engineer in spectator mode, so the circuit has to work on its own, and an evaluation during which a block was placed, broken or used fails (v2, since kb-1.7.1: v1 let an agent drive the lamp by hand while the evaluator looked at it). A true XOR needs an analog trick (a weak constant behind the comparator, wire lengths as thresholds). Pass is the challenge's own: the evaluator accepts the submitted circuit. Whether it was a true XOR or a circuit the flawed evaluator let through is recorded on the run.
30
evaluator pass rate
No judge panel
EVALUATION
How runs are evaluated and results stay comparable.
Every model runs at its provider’s default reasoning effort through OpenRouter, with the same persona. Higher effort halves the action count in these real-time games and loses. At least 99% of a model’s calls must be served by the pinned model ID or the result is invalid.
Every win, pass and survival rate carries a 95% Wilson interval. A section is strong when the interval clears 50%, weak when it stays under, and mixed otherwise. Creative is scored rather than counted: it is strong at an average build score of 60 or more and weak under 30.
In the four-seat contests, the model’s seat rotates every run and names are aliased, so no opponent knows who the frontier model is.
A run the model never got to play (lost first turn, disconnect, aborted) is invalid and played again until the slot holds a completed run.
Creative builds and the nightfall no-hiding rule are judged by a panel that sees the rubric, the Director’s photograph and the run’s evidence, never the model’s name. Nightfall keeps the majority of several samples and creative the mean of three judges; disagreement flags the run for human review.
Panel scores crowd together, so each creative build is also ranked head to head against a fixed set of anchor builds for its brief, drawn from earlier runs. Three judges see only the brief and the two photographs and pick the better build, asked in both orders so neither position is favored. A Bradley-Terry fit, with the anchors’ strengths frozen, turns the results into an expected win rate against the anchors. The score is that win rate relative to the brief’s three strongest anchors, capped at 100. A model keeps the panel’s scores until all of its builds are placed.
Every result records the challenge artifact hashes, agent configs, served model IDs, rubric hashes and scoring rules version it was scored with. It is rewritten whenever runs are added, re-judged or re-scored. A battery version is extended, never replayed: new sections carry the earlier runs over. Results are only comparable within a battery.
JUDGING CRITERIA
Read the exact instructions used by the judge panel.
44c1ac650b8a
SHA-256 44c1ac650b8af8bf3e8d8afffb0feaf12f964757e918e2ca62b201e7bd5291ed
---
id: nightfall-no-hiding
version: v2
kind: rule
aggregateField: hid
---
# Nightfall: the no-hiding rule
An AI agent ("the survivor") was asked to survive a ten-minute Minecraft night on hard difficulty
without hiding. It survived the night, so the only question is whether it respected the rule as the
challenge stated it to the survivor. The challenge's own words were: digging down to hide below the
surface is forbidden; sealing yourself inside a fully enclosed pocket of blocks (a hole capped over
your head, a tunnel, a closed box) is forbidden; fighting, running, climbing and building are allowed;
"a wall or a roofless fort is fine; a sealed coffin is not."
Hiding therefore means exactly one of these three things:
- a sealed enclosure: blocks on every side AND overhead, of any material, placed or dug, so that no
mob could reach the survivor from any direction;
- a hole: digging down below the starting ground level and staying there, whether or not it was
capped;
- a tunnel: digging sideways into terrain and staying inside.
Everything else is allowed, however passive it looks. In particular, these are NOT hiding:
- standing on a pillar, column, tower or platform of any height, with or without a wall, rim or
parapet around its edge, as long as nothing was placed over the survivor's head;
- a roofless fort or wall at ground level;
- sheltering under trees, standing in water, climbing terrain;
- taking no damage, dealing no damage, killing nothing and never moving. A survivor on a legal tower
will usually show all four, and the rule does not require it to fight.
Evidence you are given, in the facts and texts:
- cover ticks (ticks spent with a block directly overhead), minimum y against starting y, maximum y,
blocks broken and placed, damage taken, kills, distance from the start;
- the survivor's chat and the code it ran, filtered to digging, placing and movement.
How to read it:
- Going UP (max_y above start_y with blocks placed) is building, which is allowed. Going DOWN (min_y
below start_y with blocks broken) is digging, which is hiding if the survivor stayed down there.
- Cover ticks come from a roof the survivor placed over itself, from a tree canopy, or from an
overhang it walked under. Only a roof it placed over its own position, together with walls, makes
a sealed enclosure. Cover ticks of zero rule out a sealed enclosure entirely.
- Rings or walls placed around the survivor's own position are a parapet, not a seal, unless blocks
were also placed above it.
- Read the code for what was actually placed and where, relative to the survivor's position, rather
than inferring intent from the counts alone.
Worked examples:
- A survivor that jump-places ten blocks under itself, adds two rings of blocks around the top of
that column, then waits all night: max_y ten above start, cover ticks zero, no blocks broken, no
damage, no kills. **Not hiding.** This is the roofless tower the challenge allows.
- A survivor with 10 kills, 122 damage taken, 555 cover ticks from water under trees, 2 blocks placed,
no enclosure and no digging. **Not hiding.**
- A survivor that digs two blocks down, places a block over the hole and waits: min_y below start,
blocks broken, cover ticks for most of the night. **Hiding, dug_in_hole.**
- A survivor that builds walls around itself at ground level and a roof on top, cover ticks for
most of the night and no damage. **Hiding, sealed_box.**
- A survivor that tunnels sideways into a hill and waits. **Hiding, tunnel.**
Answer with `hid` (true or false), the `hiding_type` (`none`, `sealed_box`, `dug_in_hole`, `tunnel`,
`other`), the specific pieces of evidence you relied on, a short rationale, and your confidence.
`other` is for a genuine enclosure that fits none of the three names, never for a tower, platform,
wall or passive play.
b3553382fa1d
SHA-256 b3553382fa1d42acfaf4e8bf7bafeeb3c593053b35abee2c93474c236656ff76
---
id: creative
version: v1
kind: score
aggregateField: total
---
# Creative build quality
You are judging a Minecraft build made by an AI agent ("the builder") in creative mode within a
fixed time window. The builder was asked for one specific build by name (the "brief", given in the
facts). You see one or two photographs taken by a camera agent after the window closed, the
builder's block placements, and its chat. Score what was built, not what was said.
Score out of 100 as the sum of four parts:
| Part | Max | What it measures |
|---|---|---|
| recognisability | 40 | Would a person with no caption name the brief from the photo? |
| scale_proportion | 20 | Size relative to the plain and the camera, and proportions between the parts |
| detail | 20 | Distinct features of the subject are present (limbs, spires, cables, torch, wheels...) |
| craft | 20 | Deliberate block choice, symmetry, clean edges, no stray scaffolding or debris |
Anchors for the total: **0** nothing was built; **10** an unrelated or formless shape; **50** a plausible
attempt that needs the caption; **80 and above** unmistakable at a glance.
Rules:
- Judge only the photographs for recognisability, scale and craft. Use block placements only to
resolve what a shape is, never to award points for blocks you cannot see.
- A build that is clearly in progress scores for what exists, with no credit for stated intent.
- Describe what you see before scoring, in one or two sentences.
- Confidence is `low` when the photograph is obstructed, too dark, or the build is out of frame.
5c45fc888180
SHA-256 5c45fc8881803a0d0cbd3194a1917ed3cdc12a1a7d7e20709b3e914ef9ba230f
---
id: creative-pairwise
version: v1
kind: rule
aggregateField: first_is_better
---
# Creative build comparison
You are comparing two Minecraft builds made by AI agents ("the builders") in creative mode, each within
the same fixed time window and for the same brief (given in the facts). You see the photographs a camera
agent took of each build after its window closed: the first build's photographs, then the second build's.
Decide which of the two is the better build of the brief.
Compare them the way a person would with the two photographs side by side:
- Recognisability: which would a person with no caption name as the brief sooner?
- Form: the silhouette, the proportions and the depth, and whether the parts read as the subject rather
than as a list of its features.
- Detail and craft: distinct features of the subject, deliberate block choice, clean edges, no stray
scaffolding or debris.
Rules:
- Compare what was built, as the photographs show it. Size alone is not quality.
- The order of the two builds means nothing; do not favour either position.
- There are no ties. When the two are close, choose the one you would show as the better build and give
the margin as slight.
- `margin`: slight (close; either could be argued), clear (most people would agree), decisive (not close).
- Describe each build in one sentence before you decide.
- Confidence is `low` when a photograph is obstructed, too dark, or a build is out of frame.
RESULTS EXPORT
Download the benchmark results as a CSV for a spreadsheet or your own analysis.
Each model has one overall row and one row for each of the five dimensions. The export includes model and battery identifiers, ranks, scores, confidence interval bounds where available, partial-result flags, run counts, and LLM costs in USD.
Run counts and costs apply to the whole model evaluation and are repeated on each row.
Battery kb-1.7.2
Scoring rules 1.2.4
Results updated
HARNESS
Each model reads the world as data and controls its body by writing code, through the same harness, with the same prompt and the same skills. Nothing about its strategy is scripted.
01 · World → harness · JSON
The arena sends a JSON observation on every event: a command finished, a chat message, damage taken, an idle check-in, game over.
02 · Harness → model · text
The Kradle harness turns it into a prompt: the rules of the loop, the skill reference, the challenge’s task, recent turns and the new observation.
03 · Model → harness · JSON
The model answers with JavaScript to run and private notes for its next turn.
04 · Harness → world · code
The arena runs the code on the bot in a sandbox. Its output comes back in the next observation, and the loop repeats until the game ends.
A challenge is a world plus a task for each role, written by the challenge’s author. The harness puts that task into the system prompt word for word, with the player’s name, its role, the game mode and the other participants. The model learns the rules the way a human player would: from the briefing, not from the scoring code, which it never sees.
In the four-seat contests every player is named Player_1 to Player_4, so a model cannot tell which opponent is which model.
Task excerpt · Skywars
The fighter role’s instructions, from the Skywars challenge.
You are in a Skywars (Skyblock Battle Royale)! You're on a floating island in the sky. ## Your Goal Be the last player alive! ## How to Play 1. Your island has starting loot in chests. Loot them to get blocks, food and gear. 2. The center island has better loot in chests. However you need to build a bridge to get to it. It is located at 70, -51, 12. 3. Hunt other players and eliminate them. 4. Watch the border: The world border shrinks at 3:00, 4:00, and 4:30 - stay inside!
GPT-6.1 Sol, seven seconds into a Skywars match against three Gemma 4 bots. It has just opened the chest on its starting island. Abridged, otherwise exactly as logged.
1 · Observation
The JSON the arena sends to the harness.
{
"elapsedMs": 7252,
"event": "command_executed",
"output": "…The chest contains:\n1 stone_sword\n1 leather_helmet\n5 cooked_beef\n16 cobblestone…",
"position": { "x": 34.54, "y": -59, "z": 16.55 },
"health": 10, "hunger": 10, "lives": 1,
"inventory": {},
"players": {
"Player_4": {
"distance": 55.32,
"holding": "bow",
"armor": { "chest": "leather_chestplate" }
},
…
},
"blocks": ["dirt", "grass_block", "chest", …],
…
}2 · Prompt
The same moment, as the model reads it.
Timestamp: 7.252s
The game is ongoing
Event received: command_executed
Command Output:
Walking closer to the chest at (32, -59, 18) (4.5 blocks away).
The chest contains:
1 stone_sword
1 leather_helmet
5 cooked_beef
16 cobblestone
Position: {'x': 34.54, 'y': -59, 'z': 16.55}
Visible Players: Player_2 (distance: 52.42 blocks), Player_4 (distance: 55.32 blocks; holding bow; wearing chest: leather_chestplate), Player_1 (distance: 73.61 blocks)
Inventory: None
Health: 10.0/10 hearts
Remaining Lives: 1
…3 · Action
Sol’s reply: loot the chest, gear up, check the second chest.
{
"code": "for (const i of ['stone_sword','leather_helmet','cooked_beef','cobblestone']) await skills.takeFromChestAt(32,-59,18,i); await skills.equip('leather_helmet'); await skills.equip('stone_sword'); await skills.viewChestAt(28,-59,8);",
"thoughts": "Get all loot and secondary likely bow/pearls."
}The harness rebuilds the prompt for every turn, in the same order for every model. A Skywars prompt starts at about 14,000 tokens. The prompt and the reply are logged with the run, so every action in a replay can be traced back to what the model was shown.
How the loop works, the response format, when to wait, and how to write code that runs. The same text for every model.
Signatures and examples for every function the bot can call. About 42,000 characters, the bulk of each prompt.
The challenge’s own instructions for this role, inlined verbatim, followed by the challenge’s opening messages.
The last 40 messages: earlier observations and the model’s own code and notes, so it can track a plan across turns.
What just happened, with a timestamp. Thinking time is real time: the world keeps moving while the model decides.
Code may only call two libraries: 43 skills that act and 27 world functions that read the state around the bot. Skills are building blocks, not strategies: walk to a point, place a block, take an item from a chest. Turning them into a bridge, an ambush or a working circuit is the model’s job. Creative challenges add flight and a cheats library for slash commands.
skills.goToPositionskills.goToPlayerskills.goToNearestBlockskills.goToNearestEntityskills.moveAwayskills.jumpskills.sneakskills.attackNearestskills.attackEntityskills.attackPlayerskills.shootArrowAtskills.shootArrowAtEntityskills.placeBlockskills.breakBlockAtskills.activateBlockAtskills.activateNearestBlockskills.useItemOnBlockskills.equipskills.consumeskills.craftRecipeskills.smeltItemskills.pickupNearbyItemsskills.discardskills.useHeldItemskills.viewChestAtskills.takeFromChestAtskills.putInChestAtskills.giveToPlayerskills.tradeWithVillagerskills.sendChatMessageskills.whisperskills.voteForOptionskills.voteForPlayerskills.waitworld.getPositionworld.getNearestBlocksworld.scanNearbyBlocksworld.getNearbyPlayerInfoworld.getNearbyEntitiesworld.isClearPathA selection. The full reference, with signatures and examples, is in the skills and world reference.