Docs / More
Per-task medians (2.x)
Per-task token and success medians from the Prism 2.x real-agent benchmark.
Per-task medians from the 2.x benchmark. Raw sessions: 2.x.json.
Per task (medians)
| Host | Task | A tokens | B tokens | A success | B success |
|---|---|---|---|---|---|
| Claude Code | es-toolkit-15043d8e | 240,748 | 274,171 | 2/2 | 2/2 |
| Claude Code | es-toolkit-2c4c0478 | 292,002 | 238,123 | 2/2 | 2/2 |
| Claude Code | es-toolkit-323893f2 | 279,058 | 253,601 | 2/2 | 2/2 |
| Claude Code | es-toolkit-3433044c | 713,050 | 775,261 | 2/2 | 2/2 |
| Claude Code | es-toolkit-827b92f7 | 235,230 | 310,244 | 2/2 | 2/2 |
| Claude Code | es-toolkit-99af1860 | 181,074 | 145,506 | 2/2 | 2/2 |
| Claude Code | es-toolkit-d9574efd | 284,307 | 277,155 | 2/2 | 2/2 |
| Claude Code | es-toolkit-e62aa15f | 401,004 | 502,411 | 2/2 | 2/2 |
| Claude Code | hono-afb2068c | 132,792 | 138,857 | 2/2 | 2/2 |
| Claude Code | hono-c3053ccf | 118,871 | 121,722 | 2/2 | 2/2 |
| Claude Code | hono-e5bb2062 | 130,730 | 157,636 | 2/2 | 2/2 |
| opencode | es-toolkit-15043d8e | 912,158 | 441,757 | 0/1 | 1/1 |
| opencode | es-toolkit-2c4c0478 | 506,939 | 309,093 | 1/1 | 0/1 |
| opencode | es-toolkit-323893f2 | 575,588 | 249,087 | 1/1 | 0/1 |
| opencode | es-toolkit-3433044c | 1,391,775 | 1,030,587 | 0/1 | 1/1 |
| opencode | es-toolkit-827b92f7 | 619,099 | 414,169 | 0/1 | 0/1 |
| opencode | es-toolkit-99af1860 | 497,996 | 240,143 | 1/1 | 1/1 |
| opencode | es-toolkit-d9574efd | 574,282 | 540,546 | 1/1 | 1/1 |
| opencode | es-toolkit-e62aa15f | 330,705 | 680,145 | 1/1 | 1/1 |
| opencode | hono-afb2068c | 823,491 | 547,374 | 1/1 | 1/1 |
| opencode | hono-c3053ccf | 188,964 | 347,203 | 1/1 | 1/1 |
| opencode | hono-e5bb2062 | 486,401 | 413,492 | 1/1 | 1/1 |