Agentick
Agents solve procedurally generated gridworld tasks involving navigation, planning, reasoning, memory, generalization, and coordination. Scores normalize task performance between random and oracle reference policies.
The suite contains 37 tasks with several observation formats and difficulty levels. Published configurations vary in training, harness, modality, and seed count. These settings must accompany scores.
- Released
- May 2026
- Kind
- Puzzles and chess
- Level
- Everyday
- Scored
- Checked automatically
- Answers
- Agent task
- Tasks
- All public
- Measures
- Capability
- Language
- English
- Interaction
- Agentic
- Environment
- A simulator
- Items from
- Generated
- Refresh
- One fixed set
- Run by
- Anyone
- Strongest model tested
- GPT-6 Astra #4 on the capability index
- Last model added
- September 2026
- Scores refreshed
- November 20th, 2026
In the authors’ words
AI agent research spans a wide spectrum: from RL agents that learn from scratch to foundation model agents that leverage pre-trained knowledge, yet no unified benchmark enables fair comparison across these approaches. We present Agentick, a benchmark for sequential decision-making agents designed to evaluate RL, LLM, VLM, hybrid, and human agents on common ground and to power research on the fundamental challenges of sequential decision-making. Agentick provides 37 procedurally generated tasks across six capability categories, four difficulty levels, and five observation modalities, all exposed through a single Gymnasium-compatible interface. The benchmark ships with a Coding API, oracle reference policies for all tasks, pre-built SFT datasets, a composable agent harness, and a live leaderboard. An evaluation spanning 27 configurations and over 90,000 episodes reveals that no single approach dominates: GPT-5 mini leads overall at 0.309 oracle-normalized score while PPO dominates planning and multi-agent tasks; the reasoning harness multiplies LLM performance by 3-10x; and ASCII observations consistently outperform natural language. These findings highlight the substantial room for improvement that remains across all agent paradigms. Agentick's capability-decomposed, multi-modal design provides the empirical infrastructure needed to drive progress toward general autonomous agents, both as an evaluation framework and as a training ground for RL post-training of foundation models in truly sequential environments.
GPT-6 Astra leads with 0.63 among models released by September 2026 GPT-6 Astra leads with 0.63 among models released by September 2026
Best score per entry from the benchmark's own leaderboard, read on November 20th, 2026. Where a model ran with several settings or harnesses, we keep its best. Every run is in the next chart. Names in gray aren't in our model list: fine-tunes, agent systems, products and older models. They don't count toward leaders. This view shows only models we track, each at its best.
The newest model with a score here came out in September 2026, and nobody has tested a newer one. Among the 20 highest-ranked models on the capability index, these have no score yet:
- GPT-6.2 Astra
- Claude Fable 5.2
-
Gemini 4 Carbon
- Claude Opus 5.5
- and 12 more
- 0.63 GPT-6 Astra
- 0.63 Claude Opus 5
- 0.58 GPT-5.6 Sol
- 0.57 GPT-5.6 Terra
- 0.55 GPT-5.6 Luna
- 0.45 Qwen3.5-4B (SFT-250k)
- 0.35 Qwen3.5-4B (SFT-120k)
- 0.31 GPT-5 mini
- 0.29 PPO Dense (2M)
- 0.23
Qwen3.5-4B
- 0.23 PPO Dense (500k)
- 0.19
Gemini 2.5 Flash-Lite
- 0.13
Qwen3.5-2B
- 0.09
Qwen3.5-0.8B
- 0.09
Qwen3-4B
- 0.07 PPO Sparse (500k)
- 1 GPT-6 Astra ascii, high, 1 seed 0.63
- 2 Claude Opus 5 ascii, high, 1 seed 0.63
- 3 GPT-5.6 Sol ascii, high, 1 seed 0.58
- 4 GPT-5.6 Terra ascii, high, 1 seed 0.57
- 5 GPT-5.6 Luna ascii, high, 1 seed 0.55
- 6 Qwen3.5-4B (SFT-250k) ascii 0.45
- 7 Qwen3.5-4B (SFT-120k) ascii 0.35
- 8 GPT-5 mini ascii 0.31
- 9 PPO Dense (2M) rgb_array 0.29
- 10
Qwen3.5-4B ascii 0.23
- 11 PPO Dense (500k) rgb_array 0.23
- 12
Gemini 2.5 Flash-Lite Sep 2025, ascii 0.19
- 13
Qwen3.5-2B ascii 0.13
- 14
Qwen3.5-0.8B ascii 0.09
- 15
Qwen3-4B ascii 0.09
- 16 PPO Sparse (500k) rgb_array 0.07
Source: Official Agentick leaderboard
- 1 GPT-6 Astra ascii, high, 1 seed 0.63
- 2 Claude Opus 5 ascii, high, 1 seed 0.63
- 3 GPT-5.6 Sol ascii, high, 1 seed 0.58
- 4 GPT-5.6 Terra ascii, high, 1 seed 0.57
- 5 GPT-5.6 Luna ascii, high, 1 seed 0.55
- 6 GPT-5 mini ascii 0.31
- 7
Qwen3.5-4B ascii 0.23
- 8
Gemini 2.5 Flash-Lite Sep 2025, ascii 0.19
- 9
Qwen3.5-2B ascii 0.13
- 10
Qwen3.5-0.8B ascii 0.09
- 11
Qwen3-4B ascii 0.09
Source: Official Agentick leaderboard
| # | Entry | Score ± 95% CI | Setting | Released | Run on | Source |
|---|---|---|---|---|---|---|
| 1 | GPT-6 Astra | 0.63 0.5 to 0.8 | ascii, high, 1 seed, markovian_reasoner | Sep 2026 | Sep 2026 | Official leaderboard |
| 2 | Claude Opus 5 | 0.63 0.5 to 0.7 | ascii, high, 1 seed, markovian_reasoner | Jul 2026 | Sep 2026 | Official leaderboard |
| 3 | GPT-5.6 Sol | 0.58 ±0.1 | ascii, high, 1 seed, markovian_reasoner | Jul 2026 | Sep 2026 | Official leaderboard |
| 4 | GPT-5.6 Terra | 0.57 ±0.1 | ascii, high, 1 seed, markovian_reasoner | Jul 2026 | Sep 2026 | Official leaderboard |
| 5 | GPT-5.6 Luna | 0.55 ±0.1 | ascii, high, 1 seed, markovian_reasoner | Jul 2026 | Sep 2026 | Official leaderboard |
| 6 | Qwen3.5-4B (SFT-250k) open weights | 0.45 0.4 to 0.5 | ascii, markovian_zero_shot | May 2026 | Official leaderboard | |
| 7 | Qwen3.5-4B (SFT-120k) open weights | 0.35 0.3 to 0.4 | ascii, markovian_zero_shot | May 2026 | Official leaderboard | |
| 8 | GPT-5 mini | 0.31 | ascii, MarkovianReasoner | Aug 2025 | Mar 2026 | Official leaderboard |
| 9 | PPO Dense (2M) open weights | 0.29 0.2 to 0.4 | rgb_array | Mar 2026 | Official leaderboard | |
| 10 | | 0.23 0.2 to 0.3 | ascii, markovian_reasoner | Feb 2026 | Mar 2026 | Official leaderboard |
| 11 | PPO Dense (500k) open weights | 0.23 ±0.1 | rgb_array | Mar 2026 | Official leaderboard | |
| 12 | | 0.19 ±0.1 | ascii, markovian_reasoner | Sep 2025 | Mar 2026 | Official leaderboard |
| 13 | | 0.13 ±0.1 | ascii, markovian_reasoner | Feb 2026 | Mar 2026 | Official leaderboard |
| 14 | | 0.09 ±0 | ascii, markovian_reasoner | Feb 2026 | Mar 2026 | Official leaderboard |
| 15 | | 0.09 | ascii, markovian_reasoner | Apr 2025 | Mar 2026 | Official leaderboard |
| 16 | PPO Sparse (500k) open weights | 0.07 ±0 | rgb_array | Mar 2026 | Official leaderboard |
| # | Model | Score ± 95% CI | Setting | Released | Run on | Source |
|---|---|---|---|---|---|---|
| 1 | GPT-6 Astra | 0.63 0.5 to 0.8 | ascii, high, 1 seed, markovian_reasoner | Sep 2026 | Sep 2026 | Official leaderboard |
| 2 | Claude Opus 5 | 0.63 0.5 to 0.7 | ascii, high, 1 seed, markovian_reasoner | Jul 2026 | Sep 2026 | Official leaderboard |
| 3 | GPT-5.6 Sol | 0.58 ±0.1 | ascii, high, 1 seed, markovian_reasoner | Jul 2026 | Sep 2026 | Official leaderboard |
| 4 | GPT-5.6 Terra | 0.57 ±0.1 | ascii, high, 1 seed, markovian_reasoner | Jul 2026 | Sep 2026 | Official leaderboard |
| 5 | GPT-5.6 Luna | 0.55 ±0.1 | ascii, high, 1 seed, markovian_reasoner | Jul 2026 | Sep 2026 | Official leaderboard |
| 6 | GPT-5 mini | 0.31 | ascii, MarkovianReasoner | Aug 2025 | Mar 2026 | Official leaderboard |
| 7 | | 0.23 0.2 to 0.3 | ascii, markovian_reasoner | Feb 2026 | Mar 2026 | Official leaderboard |
| 8 | | 0.19 ±0.1 | ascii, markovian_reasoner | Sep 2025 | Mar 2026 | Official leaderboard |
| 9 | | 0.13 ±0.1 | ascii, markovian_reasoner | Feb 2026 | Mar 2026 | Official leaderboard |
| 10 | | 0.09 ±0 | ascii, markovian_reasoner | Feb 2026 | Mar 2026 | Official leaderboard |
| 11 | | 0.09 | ascii, markovian_reasoner | Apr 2025 | Mar 2026 | Official leaderboard |
Rankings by part
We read 6 parts from the benchmark’s own leaderboard, each ranking models on one part of the test, such as Generalization, Memory and Multi agent. Claude Opus 5 leads 4 of them, and 2 other models lead the rest. Arrows show how far a model moved from its overall rank among the models we track.
Generalization
11 models we track have a score here.
- 1 Claude Opus 5 ↑1 0.58
- 2 GPT-6 Astra ↓1 0.58
- 3 GPT-5.6 Luna ↑2 0.50
- 4 GPT-5.6 Terra 0.50
- 5 GPT-5 mini ↑1 0.44
- 6 GPT-5.6 Sol ↓3 0.42
- 7
Qwen3.5-4B 0.34
- 8
Gemini 2.5 Flash-Lite Sep 2025 0.29
- 9
Qwen3.5-2B 0.17
- 10
Qwen3.5-0.8B 0.14
Memory
11 models we track have a score here.
- 1 Claude Opus 5 ↑1 0.50
- 2 GPT-5.6 Luna ↑3 0.44
- 3 GPT-5.6 Sol 0.44
- 4 GPT-5.6 Terra 0.44
- 5 GPT-6 Astra ↓4 0.38
- 6 GPT-5 mini 0.35
- 7
Qwen3.5-4B 0.29
- 8
Qwen3.5-2B ↑1 0.21
- 9
Gemini 2.5 Flash-Lite Sep 2025 ↓1 0.16
- 10
Qwen3-4B ↑1 0.15
Multi agent
11 models we track have a score here.
- 1 GPT-5.6 Luna ↑4 0.50
- 2 GPT-5.6 Terra ↑2 0.45
- 3 GPT-6 Astra ↓2 0.45
- 4 GPT-5.6 Sol ↓1 0.40
- 5 Claude Opus 5 ↓3 0.40
- 6 GPT-5 mini 0.15
- 7
Qwen3.5-4B 0.13
- 8
Gemini 2.5 Flash-Lite Sep 2025 0.10
- 9
Qwen3.5-2B 0.05
- 10
Qwen3-4B ↑1 0.04
Navigation
11 models we track have a score here.
- 1 Claude Opus 5 ↑1 0.78
- 2 GPT-6 Astra ↓1 0.78
- 3 GPT-5.6 Terra ↑1 0.75
- 4 GPT-5.6 Sol ↓1 0.72
- 5 GPT-5.6 Luna 0.69
- 6 GPT-5 mini 0.46
- 7
Gemini 2.5 Flash-Lite Sep 2025 ↑1 0.24
- 8
Qwen3.5-4B ↓1 0.22
- 9
Qwen3.5-2B 0.14
- 10
Qwen3-4B ↑1 0.07
Planning
11 models we track have a score here.
- 1 GPT-6 Astra 0.83
- 2 GPT-5.6 Sol ↑1 0.78
- 3 Claude Opus 5 ↓1 0.72
- 4 GPT-5.6 Luna ↑1 0.69
- 5 GPT-5.6 Terra ↓1 0.69
- 6 GPT-5 mini 0.33
- 7
Qwen3.5-4B 0.31
- 8
Gemini 2.5 Flash-Lite Sep 2025 0.25
- 9
Qwen3.5-2B 0.24
- 10
Qwen3.5-0.8B 0.16
Reasoning
11 models we track have a score here.
- 1 Claude Opus 5 ↑1 0.78
- 2 GPT-6 Astra ↓1 0.78
- 3 GPT-5.6 Sol 0.72
- 4 GPT-5.6 Terra 0.56
- 5 GPT-5.6 Luna 0.50
- 6 GPT-5 mini 0.13
- 7
Qwen3.5-4B 0.12
- 8
Gemini 2.5 Flash-Lite Sep 2025 0.09
- 9
Qwen3.5-2B 0.05
- 10
Qwen3.5-0.8B 0.02
Scores at each effort level
Models the leaderboard ran at two or more effort levels. We don’t have enough cost data here to say what the extra effort costs.
Source: Official Agentick leaderboard
| Model | Effort | Harness | Score | Cost per task | Output tokens |
|---|---|---|---|---|---|
| GPT-6 Astra | ascii, high, 1 seed | markovian_reasoner | 0.63 | ||
| Claude Opus 5 | ascii, high, 1 seed | markovian_reasoner | 0.63 | ||
| GPT-5.6 Sol | ascii, high, 1 seed | markovian_reasoner | 0.58 | ||
| GPT-5.6 Terra | ascii, high, 1 seed | markovian_reasoner | 0.57 | ||
| GPT-5.6 Luna | ascii, high, 1 seed | markovian_reasoner | 0.55 | ||
| GPT-5 mini | ascii | MarkovianReasoner | 0.31 | ||
| ascii | markovian_reasoner | 0.23 | |||
| ascii | markovian_reasoner | 0.19 | |||
| language | markovian_reasoner | 0.18 | |||
| ascii | markovian_reasoner | 0.13 | |||
| language | markovian_reasoner | 0.12 | |||
| ascii | markovian_reasoner | 0.09 | |||
| rgb_array | markovian_zero_shot | 0.09 | |||
| ascii | markovian_reasoner | 0.09 | |||
| language | markovian_zero_shot | 0.06 | |||
| ascii | markovian_zero_shot | 0.06 | |||
| language | markovian_reasoner | 0.06 | |||
| ascii | markovian_zero_shot | 0.05 | |||
| language | markovian_reasoner | 0.05 | |||
| language | markovian_zero_shot | 0.03 | |||
| ascii | markovian_zero_shot | 0.02 | |||
| ascii | markovian_zero_shot | 0.02 | |||
| ascii | markovian_zero_shot | 0.02 | |||
| language | markovian_zero_shot | 0.02 | |||
| language | markovian_zero_shot | 0.02 | |||
| language | markovian_zero_shot | 0.02 |
Scores at each effort level
Models the leaderboard ran at two or more effort levels. We don’t have enough cost data here to say what the extra effort costs.
Source: Official Agentick leaderboard
| Model | Effort | Harness | Score | Cost per task | Output tokens |
|---|---|---|---|---|---|
| GPT-6 Astra | ascii, high, 1 seed | markovian_reasoner | 0.63 | ||
| Claude Opus 5 | ascii, high, 1 seed | markovian_reasoner | 0.63 | ||
| GPT-5.6 Sol | ascii, high, 1 seed | markovian_reasoner | 0.58 | ||
| GPT-5.6 Terra | ascii, high, 1 seed | markovian_reasoner | 0.57 | ||
| GPT-5.6 Luna | ascii, high, 1 seed | markovian_reasoner | 0.55 | ||
| GPT-5 mini | ascii | MarkovianReasoner | 0.31 | ||
| ascii | markovian_reasoner | 0.23 | |||
| ascii | markovian_reasoner | 0.19 | |||
| language | markovian_reasoner | 0.18 | |||
| ascii | markovian_reasoner | 0.13 | |||
| language | markovian_reasoner | 0.12 | |||
| ascii | markovian_reasoner | 0.09 | |||
| rgb_array | markovian_zero_shot | 0.09 | |||
| ascii | markovian_reasoner | 0.09 | |||
| language | markovian_zero_shot | 0.06 | |||
| ascii | markovian_zero_shot | 0.06 | |||
| language | markovian_reasoner | 0.06 | |||
| ascii | markovian_zero_shot | 0.05 | |||
| language | markovian_reasoner | 0.05 | |||
| language | markovian_zero_shot | 0.03 | |||
| ascii | markovian_zero_shot | 0.02 | |||
| ascii | markovian_zero_shot | 0.02 | |||
| ascii | markovian_zero_shot | 0.02 | |||
| language | markovian_zero_shot | 0.02 | |||
| language | markovian_zero_shot | 0.02 | |||
| language | markovian_zero_shot | 0.02 |
4 models have held number one on the official leaderboard since its release in May 2026
Lab claims count from the day the card came out, independent runs from the model’s release.
Source: Official Agentick leaderboard
Scores over time
All the scores we have, one dot each, so a model tested at several settings appears more than once.
Source: Official Agentick leaderboard
No lab model card reports it yet
We read the headline results of 540 launch posts from 57 labs, and none of them list this benchmark.
More in Games and long tasks
- ARC-AGI-3 In use Interactive game environments where agents must explore, infer goals and rules, and act efficiently without instructions. GPT-6.2 Astra 100%
- Orak Fading LLM agents play 12 popular video games across most genres through an MCP interface. Each game scores in-game progress or outcomes, and the board ranks models by their average rank across the games.
Gemini 2.5 Pro 3.90
- MineDojo Fading Agents complete open-ended tasks in Minecraft, such as mining blocks and fighting mobs, from a large simulation suite. Programmatic checks score many tasks.
Seed2.0 Pro 49%
- TextQuests Only in system cards Agents play 25 classic interactive fiction games. The score is harmful actions per 100 steps, where lower is better. The board also reports game progress.
Llama 4 Scout 7.00
- GameWorld Only in system cards Multimodal game agents play 34 browser games across 170 tasks. Outcomes are checked from the game state. GPT-6.1 Terra 76.8%
- Swarm Bench Only in system cards Models orchestrate agent swarms on complex tasks that benefit from coordinated decomposition and parallel execution. Kimi K3 76.3%
Data audit
All 9 checks pass. Every benchmark
Catalog record 2 pass
- Pass Task access All tasks are public.
- Pass Confidence Checked against a primary source.
Scores 3 pass 2 notes
- Pass Independent runs Scores from the official leaderboard (16 entries).
- Pass Models scored 11 models we track, plus 5 other entries on its leaderboard.
- Pass Newest score The newest is from September 2026, for GPT-6 Astra.
- Note Frontier models Only 1 of the top 10 has a score. The strongest tested is GPT-6 Astra (#4).
- Note Headroom Scored as a plain number, with no fixed ceiling.
Leaderboard import 3 pass
- Pass Importer We read the board on roger-creus.github.io: 26 rows for models we track and 5 other entries at the last read.
- Pass Last read November 20th, 2026.
- Pass Read worked The last read worked.
Consistency 1 pass 1 note
- Note Lab claims No model has both a lab number and an independent run.
- Pass Known issues No published audit found problems.