SpaceChem Bench
Agents program reactors to assemble molecules in SpaceChem. The main ranking measures how close their solutions come to the community’s records.
Wincon ranks 16 models on 55 puzzles. Each model gets one timed attempt per puzzle, including 18 puzzles that require laying out production lines.
- Made by
- Wincon Research
- Released
- November 2026
- Kind
- Puzzles and chess
- Level
- Competition
- Scored
- Checked automatically
- Answers
- Agent task
- Tasks
- All public
- Measures
- Capability
- Language
- English
- Interaction
- Agentic
- Environment
- SpaceChem
- Items from
- Real work and use
- Refresh
- One fixed set
- Run by
- Anyone
- Strongest model tested
- GPT-6 Astra #4 on the capability index
- Last model added
- October 2026
- Scores refreshed
- November 20th, 2026
GPT-6 Astra leads with 74.3
Best score per model from the benchmark's own leaderboard, read on November 20th, 2026. Where a model ran with several settings or harnesses, we keep its best.
- 74.3 GPT-6 Astra
- 69.7 GPT-6.1 Sol
- 66.2 Claude Opus 5.5
- 65.2
Gemini 4 Argon
- 53.7 Claude Sonnet 5.5
- 49.8 Claude Fable 5.1
- 42.6
Gemini 3.8 Flash
- 41.7 Kimi K3.1 Pro
- 36.8
Muse Wonder
- 31.8
DeepSeek V4.1 Flash
- 29.4
Muse Spark 1.3
- 29.4 Grok 4.7
- 24.5 Claude Haiku 5.5
- 19.3 GPT-6 Luna
- 16.4 GLM-5.3
- 9.80 MiMo-V2.6 Pro
- 1 GPT-6 Astra 74.3
- 2 GPT-6.1 Sol 69.7
- 3 Claude Opus 5.5 66.2
- 4
Gemini 4 Argon 65.2
- 5 Claude Sonnet 5.5 53.7
- 6 Claude Fable 5.1 49.8
- 7
Gemini 3.8 Flash 42.6
- 8 Kimi K3.1 Pro 41.7
- 9
Muse Wonder 36.8
- 10
DeepSeek V4.1 Flash 31.8
- 11
Muse Spark 1.3 29.4
- 12 Grok 4.7 29.4
- 13 Claude Haiku 5.5 24.5
- 14 GPT-6 Luna 19.3
- 15 GLM-5.3 16.4
- 16 MiMo-V2.6 Pro 9.80
Source: Official SpaceChem Bench leaderboard
| # | Model | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | GPT-6 Astra | 74.3 | Sep 2026 | Official leaderboard | |
| 2 | GPT-6.1 Sol | 69.7 | Sep 2026 | Official leaderboard | |
| 3 | Claude Opus 5.5 | 66.2 | Sep 2026 | Official leaderboard | |
| 4 | | 65.2 | Sep 2026 | Official leaderboard | |
| 5 | Claude Sonnet 5.5 | 53.7 | Sep 2026 | Official leaderboard | |
| 6 | Claude Fable 5.1 | 49.8 | Sep 2026 | Official leaderboard | |
| 7 | | 42.6 | Sep 2026 | Official leaderboard | |
| 8 | Kimi K3.1 Pro | 41.7 | Oct 2026 | Official leaderboard | |
| 9 | | 36.8 | Oct 2026 | Official leaderboard | |
| 10 | | 31.8 | Sep 2026 | Official leaderboard | |
| 11 | | 29.4 | Sep 2026 | Official leaderboard | |
| 12 | Grok 4.7 | 29.4 | Sep 2026 | Official leaderboard | |
| 13 | Claude Haiku 5.5 | 24.5 | Oct 2026 | Official leaderboard | |
| 14 | GPT-6 Luna | 19.3 | Sep 2026 | Official leaderboard | |
| 15 | GLM-5.3 | 16.4 | Aug 2026 | Official leaderboard | |
| 16 | MiMo-V2.6 Pro | 9.80 | Sep 2026 | Official leaderboard |
Number one over time
Lab claims count from the day the card came out, independent runs from the model’s release.
Source: Official SpaceChem Bench 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 SpaceChem Bench 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 10 checks pass. Every benchmark
Catalog record 2 pass
- Pass Task access All tasks are public.
- Pass Confidence Checked against a primary source.
Scores 4 pass 1 note
- Pass Independent runs Scores from the official leaderboard (16 entries).
- Pass Models scored 16 models we track.
- Pass Newest score The newest is from October 2026, for Kimi K3.1 Pro.
- Pass Frontier models 6 of the top 10 have a score, led by 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 spacechem.wincon.dev: 16 rows for models we track 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.