Simulation gallery
| Model | Harness | Agents | Setting | Sessions | Hours | Tokens (B) | Scarry | Open simulation |
|---|---|---|---|---|---|---|---|---|
| World-1 collaboration | ||||||||
| claude-opus-5 | Claude Code | 6 | default | 596 | 198 | 3.96 | 1.159 | Open simulation → |
| gpt-5.6-sol | Pi | 6 | default | 606 | 75 | 0.69 | 1.059 | Open simulation → |
| gpt-6-astra | Pi | 6 | default | 607 | 18 | 0.08 | 1.005 | Open simulation → |
| gemini-3.8-flash | Pi | 6 | default | 604 | 272 | 11.73 | 1.251 | Open simulation → |
| kimi-k2.7-code | Pi | 6 | default | 605 | 109 | 1.60 | 1.105 | |
| glm-5.3 | Pi | 6 | default | 605 | 82 | 0.50 | 1.038 | Open simulation → |
| World-2 collaboration | ||||||||
| claude-opus-5 | Claude Code | 6 | default | 547 | 324 | 5.90 | 1.023 | Open simulation → |
| gpt-5.6-sol | Pi | 6 | default | 1,205 | 132 | 1.17 | 1.019 | Open simulation → |
| gpt-6-astra | Pi | 6 | default | 606 | 22 | 0.14 | 1.000 | Open simulation → |
| gemini-3.8-flash | Pi | 6 | default | 604 | 282 | 15.70 | 1.144 | Open simulation → |
| kimi-k2.7-code | Pi | 6 | default | 606 | 162 | 2.81 | 1.045 | |
| glm-5.3 | Pi | 6 | default | 603 | 60 | 0.41 | 1.006 | |
| World-3 collaboration | ||||||||
| claude-opus-5 | Claude Code | 26 | default | 1,947 | 319 | 4.54 | 1.281 | |
| gpt-5.6-sol | Codex | 26 | default | 877 | 113 | 2.18 | 1.060 | Open simulation → |
| gpt-6-astra | Pi | 26 | default | 1,319 | 97 | 0.44 | 1.017 | |
| gemini-3.8-flash | Pi | 26 | default | 746 | 154 | 5.57 | 1.305 | |
| kimi-k2.7-code | Kimi CLI | 26 | default | 1,252 | 253 | 5.49 | 1.124 | Open simulation → |
| glm-5.3 | Pi | 26 | default | 1,649 | 148 | 1.15 | 1.191 | |
| Discovery prompt, World-1 | ||||||||
| claude-fable-5-1 | Claude Code | 6 | default | 3,003 | 120 | 0.73 | 1.021 | |
| claude-fable-5-1 | Claude Code | 6 | discovery prompt | 2,972 | 180 | 1.44 | 1.060 | |
| claude-opus-5-5 | Claude Code | 6 | default | 3,007 | 78 | 0.41 | 1.055 | |
| claude-opus-5-5 | Claude Code | 6 | discovery prompt | 3,001 | 109 | 0.68 | 1.092 | |
| gpt-6-astra | Pi | 6 | discovery prompt | 605 | 54 | 0.20 | 1.003 | Open simulation → |
| Communication, World-1 | ||||||||
| claude-opus-5 | Claude Code | 6 | isolated | 601 | 89 | 1.72 | 1.135 | Open simulation → |
| gpt-5.6-sol | Pi | 6 | isolated | 600 | 88 | 0.69 | 1.053 | |
| gpt-6-astra | Pi | 6 | isolated | 604 | 74 | 0.40 | 1.006 | Open simulation → |
| gemini-3.8-flash | Pi | 6 | isolated | 605 | 178 | 4.37 | 1.194 | Open simulation → |
| kimi-k2.7-code | Kimi CLI | 6 | isolated | 600 | 119 | 1.71 | 1.108 | Open simulation → |
| gpt-5.6-sol | Pi | 6 | forced, ≥1 of 10 sessions | 605 | 76 | 0.51 | 1.051 | |
| gpt-5.6-sol | Pi | 6 | forced, ≥3 of 10 sessions | 606 | 80 | 0.56 | 1.048 | |
| gpt-5.6-solclaude-opus-5 | PiClaude Code | 4 + 2 | mixed models | 604 | 115 | 1.08 | 1.120 | |
| Ownership, World-1 | ||||||||
| claude-opus-5 | Claude Code | 1 | one agent, all packages | 135 | 86 | 1.49 | 1.089 | Open simulation → |
| claude-opus-5 | Claude Code | 4 | unassigned | 555 | 191 | 3.55 | 1.098 | Open simulation → |
| claude-opus-5 | Claude Code | 6 | unassigned | 516 | 221 | 3.57 | 1.089 | Open simulation → |
| claude-opus-5 | Claude Code | 10 | unassigned | 612 | 228 | 3.63 | 1.096 | Open simulation → |
| gpt-5.6-sol | Pi | 1 | one agent, all packages | 431 | 78 | 0.72 | 1.077 | Open simulation → |
| gemini-3.8-flash | Pi | 1 | one agent, all packages | 339 | 155 | 6.22 | 1.204 | Open simulation → |
| kimi-k2.7-code | Pi | 1 | one agent, all packages | 600 | 98 | 2.60 | 1.122 | |
| Budget sharing, World-1 | ||||||||
| claude-opus-5 | Claude Code | 6 | budget-aware prompt | 483 | 197 | 3.60 | 1.172 | Open simulation → |
| claude-opus-5 | Claude Code | 6 | shared budget, default | 277 | 74 | 1.15 | 1.107 | |
| claude-opus-5 | Claude Code | 6 | shared budget, guided | 213 | 65 | 0.94 | 1.100 | Open simulation → |
| gpt-5.6-sol | Pi | 6 | budget-aware prompt | 600 | 57 | 0.46 | 1.054 | Open simulation → |
| gpt-5.6-sol | Pi | 6 | shared budget, default | 372 | 32 | 0.20 | 1.014 | |
| gpt-5.6-sol | Pi | 6 | shared budget, guided | 594 | 51 | 0.42 | 1.042 | Open simulation → |
| gemini-3.8-flash | Pi | 6 | budget-aware prompt | 548 | 233 | 10.18 | 1.207 | |
| gemini-3.8-flash | Pi | 6 | shared budget, default | 493 | 210 | 8.00 | 1.250 | |
| gemini-3.8-flash | Pi | 6 | shared budget, guided | 596 | 209 | 9.65 | 1.194 | |
| Total | 47 runs | 40,361 | 6,367 | 134.96 | ||||
Other simulations
| Model | Harness | Agents | Setting | Sessions | Hours | Tokens (B) | Scarry | Open simulation |
|---|---|---|---|---|---|---|---|---|
| World-3 | ||||||||
| claude-sonnet-5 | claude-code | 26 | default | 720 | 93 | 3.13 | — | Open simulation → |