Traces

Every run: the agent's full trajectory, its submissions over the 20-hour budget, and the final score.

Git to Zig

task page →50 runs · indexed 2026-09-02

#	Model	Reward	Cost	In tok	Out tok	Time	Steps	Submits
1	Claude Fable 5.1	0.5599	$663.30	1199.7M	2.6M	18h 57m	2674	17
2	Claude Fable 5.1	0.5146	$566.08	1019.0M	2.3M	19h 11m	2360	13
3	Claude Fable 5.1	0.5049	$432.88	762.8M	2.1M	18h 58m	1732	13
4	Claude Fable 5.1	0.4886	$363.29	631.2M	2.0M	18h 58m	1562	11
5	Claude Fable 5.1	0.4844	$390.64	686.7M	2.0M	19h 13m	1611	26
6	GPT-5.6	0.3440	$1238.08	1225.0M	944k	19h 8m	2627	21
7	GPT-5.6	0.2804	$1082.72	1080.1M	850k	19h 36m	2314	42
8	GPT-5.6	0.2801	$1175.97	1171.4M	884k	19h 37m	2602	29
9	GPT-5.6	0.2369	$955.56	945.7M	722k	19h 4m	1982	13
10	GPT-5.6	0.1874	$770.31	763.5M	629k	19h 0m	1668	6
11	Gemini 3.7 Flash	0.1808	$157.01	1517.8M	2.4M	19h 43m	3256	1
12	Kimi K3	0.1768	$251.00	752.4M	1.2M	20h 2m	1619	1
13	Kimi K3	0.1733	$200.20	602.7M	965k	19h 43m	1483	13
14	Grok 4.6	0.1693	$652.70	657.8M	2.2M	19h 59m	2502	122
15	Grok 4.6	0.1618	$614.44	610.4M	2.6M	18h 0m	2324	73
16	Kimi K3	0.1616	$254.18	761.8M	1.1M	19h 47m	1668	9
17	Kimi K3	0.1574	$259.84	779.9M	1.1M	19h 53m	1765	2
18	GLM-5.3	0.1556	$225.36	842.0M	964k	19h 45m	2012	17
19	GLM-5.3	0.1531	$195.54	692.6M	689k	19h 54m	1480	2
20	Grok 4.6	0.1516	$806.68	819.7M	2.2M	19h 59m	2981	186
21	Inkling	0.1510	$17.77	18.6M	32k	1h 29m	255	1
22	Gemini 3.7 Flash	0.1495	$79.59	838.0M	1.2M	20h 0m	1855	2
23	Kimi K3	0.1473	$222.64	660.4M	1.1M	19h 44m	1491	1
24	Gemini 3.7 Flash	0.1433	$101.57	1069.8M	1.5M	19h 58m	2372	1
25	GLM-5.3	0.1390	$212.77	708.2M	852k	19h 49m	1640	49
26	GLM-5.3	0.1378	$248.79	932.5M	906k	19h 55m	2219	21
27	Grok 4.6	0.1352	$482.88	483.0M	1.6M	11h 22m	1767	84
28	DeepSeek V4 Flash Vision Exp	0.1304	$28.18	1759.3M	2.2M	19h 24m	3723	15
29	Grok 4.6	0.1269	$408.61	405.7M	1.8M	19h 55m	1558	48
30	Gemini 3.7 Flash	0.1237	$71.55	746.3M	1.3M	19h 57m	1690	4
31	GLM-5.3	0.1219	$188.86	712.0M	621k	19h 43m	1318	16
32	DeepSeek V4 Flash Vision Exp	0.1213	$28.06	1756.1M	2.2M	19h 43m	3793	11
33	Qwen3.8-Max	0.1212	$184.24	696.5M	1.3M	19h 33m	1805	12
34	DeepSeek V4 Flash Vision Exp	0.1193	$26.16	1630.6M	2.1M	19h 37m	3584	8
35	DeepSeek V4 Flash Vision Exp	0.1079	$21.83	1364.0M	1.7M	17h 17m	2977	8
36	Qwen3.8-Max	0.1038	$196.74	742.9M	1.4M	19h 38m	1831	26
37	Qwen3.8-Max	0.1023	$180.44	685.7M	1.2M	19h 14m	1776	2
38	DeepSeek V4 Flash Vision Exp	0.0988	$20.50	1272.8M	1.7M	19h 17m	2732	11
39	Qwen3.8-Max	0.0987	$178.20	662.0M	1.7M	19h 50m	1710	6
40	Qwen3.8-Max	0.0927	$190.64	725.2M	1.3M	19h 52m	1620	35
41	Muse Spark 1.2	0.0185	$42.33	250.7M	911k	3h 1m	578	5
42	Inkling	0.0093	$17.52	18.3M	28k	1h 29m	208	3
43	Inkling	0.0019	$16.36	17.1M	30k	17m	280	1
44	Muse Spark 1.2	0.0000†	$2.45	9.3M	123k	59m	86	2
45	Inkling	0.0000	$3.15	3.3M	13k	5m	86	2
46	Muse Spark 1.2	0.0000†	$5.53	26.0M	135k	1h 56m	175	4
47	Muse Spark 1.2	0.0000†	$59.99	358.3M	658k	4h 22m	779	1
48	Gemini 3.7 Flash	0.0000†	$76.41	819.6M	1.5M	19h 43m	1734	1
49	Inkling	0.0000†	$5.57	5.8M	17k	1h 55m	118	2
50	Muse Spark 1.2	0.0000†	$2.14	7.7M	83k	1h 10m	93	3

† score adjusted by manual review; the trace's Judge tab carries the measured score and the rationale.