Traces

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

Multi-GPU Efficient Finetuning

task page →110 runs · indexed 2026-10-08
#ModelRewardCostIn tokOut tokTimeStepsSubmits
1Inkling0.5238$11.7812.1M59k1h 35m1661
2Claude Opus 5.50.4286$18.5551.6M316k17h 25m2014
3Muse Spark 1.20.3810$64.94126.7M662k16h 40m38627
4Claude Sonnet 5.50.3571$29.39111.1M559k18h 36m2642
5DeepSeek V4 Flash Vision Exp0.3333$0.8236.7M207k19h 58m2381
6Kimi K30.3333$19.1518.0M131k19h 9m1676
7Claude Opus 5.50.3333$13.0335.0M224k18h 31m1795
8Claude Opus 5.50.3095$18.4950.1M325k18h 23m1925
9Muse Spark 1.20.3095$11.9541.8M358k12h 44m19924
10Muse Spark 1.30.3095$25.5021.9M121k19h 28m1972
11Kimi K30.3095$16.5615.3M148k19h 10m1343
12Gemini 4 Argon0.3095$59.0860.8M237k16h 56m1985
13Claude Fable 50.2857$23.6614.1M128k18h 18m1192
14GLM-5.30.2857$12.8826.8M144k19h 0m2342
15Claude Fable 5.10.2857$5.657.6M75k17h 4m982
16Claude Opus 5.50.2857$21.9461.0M375k16h 29m2073
17Qwen3.8-Max0.2857$6.4217.1M316k18h 44m2135
18Claude Fable 5.10.2857$5.387.9M60k16h 16m1183
19Claude Haiku 5.50.2857$3.4948.9M402k16h 53m1633
20GLM-5.3 Flash0.2857$0.6810.0M99k15h 21m1243
21GLM-5.30.2857$10.4218.4M127k19h 10m1756
22Claude Opus 50.2857$20.6727.0M193k18h 45m1775
23Claude Haiku 5.50.2619$1.7824.5M255k17h 19m1352
24Claude Opus 5.50.2619$17.3345.0M317k18h 13m1784
25GLM-5.3 Flash0.2619$1.1715.8M143k16h 53m1343
26Claude Fable 50.2619$17.729.8M108k18h 5m1064
27Muse Spark 1.20.2619$355.312070.9M1.9M15h 46m592022
28GLM-5.3 Flash0.2619$0.9712.5M122k17h 37m1272
29Claude Sonnet 5.50.2619$34.13135.6M549k19h 19m3382
30Claude Fable 50.2619$30.0318.9M149k19h 3m1525
31Claude Fable 5.10.2619$6.128.6M75k18h 44m1124
32Qwen3.8-Max0.2619$6.1516.6M296k17h 38m22010
33Claude Opus 50.2381$22.4128.9M215k18h 43m1772
34GPT-5.6 Sol0.2381$19.3323.5M110k15h 47m1673
35Claude Fable 50.2381$19.0711.3M104k19h 17m1208
36GLM-5.3 Flash0.2381$1.3629.9M170k18h 53m2192
37Claude Opus 50.2381$21.9128.5M201k18h 44m1805
38Claude Fable 50.2381$17.669.8M106k18h 38m1066
39Claude Haiku 5.50.2143$2.4634.8M304k17h 30m1583
40Gemini 4 Argon0.2143$71.7765.5M249k18h 37m2008
41Claude Fable 5.10.2143$5.267.1M71k18h 19m923
42Gemini 3.8 Flash0.2143$4.9123.2M78k4h 58m1866
43GPT-6 Astra0.2143$155.5978.1M233k19h 27m2996
44Claude Opus 50.2143$16.3319.4M178k17h 33m1423
45Gemini 3.7 Flash0.2143$10.2429.6M72k9h 49m18131
46Claude Haiku 5.50.1905$3.2445.3M385k19h 9m1571
47GPT-6 Astra0.1905$229.70113.1M288k19h 53m3646
48Muse Spark 1.20.1905$5.5220.2M214k2h 15m1353
49Gemini 4 Argon0.1905$119.5980.8M260k19h 19m2303
50GPT-6 Astra0.1905$125.1064.1M212k19h 30m2616
51Kimi K30.1905$16.1311.8M128k18h 47m1161
52Gemini 3.7 Flash0.1667$11.8531.1M102k19h 53m20210
53Claude Sonnet 5.50.1667$34.90136.3M593k18h 28m3143
54Claude Sonnet 5.50.1667$37.45150.3M578k14h 47m3433
55Claude Fable 5.10.1667$6.929.6M88k17h 1m1014
56DeepSeek V4 Flash Vision Exp0.1429$0.9241.7M226k19h 50m2512
57Muse Spark 1.30.1429$16.1814.5M230k20h 0m1591
58Gemini 4 Argon0.1429$59.3568.3M219k18h 18m2077
59Gemini 4 Argon0.1190$99.1676.3M304k18h 34m2334
60Gemini 3.8 Flash0.0952$10.1937.4M109k8h 17m22711
61GPT-6 Astra0.0714$291.48141.3M320k19h 38m4026
62Claude Haiku 5.50.0714$4.8271.2M466k18h 3m2191
63GLM-5.30.0714$17.6133.2M221k18h 56m2112
64Qwen3.8-Max-09020.0476$21.6951.3M683k18h 46m3343
65Muse Spark 1.20.0476$35.64132.6M581k10h 53m45113
66Qwen3.8-Max-09020.0476$21.7463.3M874k19h 49m3973
67Kimi K30.0476$25.1335.1M233k19h 37m2361
68Inkling0.0476$5.355.5M24k1h 26m1322
69GPT-5.6 Sol0.0238$19.6525.0M71k12h 53m1333
70Inkling0.0238$14.2214.8M45k2h 19m2191
71DeepSeek V4 Flash Vision Exp0.0238$0.6629.1M168k19h 41m2262
72Grok 4.70.0000$60.2136.3M189k18h 34m1842
73Muse Spark 1.30.0000$22.9620.6M169k19h 58m1575
74Qwen3.8-Max0.0000$4.6511.0M287k19h 49m1835
75Gemini 3.7 Flash0.0000$22.1847.8M107k18h 31m2992
76Grok 4.70.0000$37.8924.1M164k18h 10m1402
77GPT-5.6 Sol0.0000$16.4525.9M78k16h 50m1802
78GLM-5.30.0000$20.2634.9M208k19h 31m2114
79Claude Opus 50.0000$20.7027.2M184k19h 13m1733
80Inkling0.0000$6.666.8M46k2h 1m1351
81GPT-5.6 Sol0.0000$24.1334.7M95k16h 10m1812
82GPT-5.6 Sol0.0000$30.4139.1M90k14h 8m1993
83Grok 4.60.0000$47.0350.4M373k19h 22m26113
84Qwen3.8-Max0.0000$3.679.9M171k19h 13m1533
85Qwen3.8-Max-09020.0000$6.948.9M369k5h 21m1201
86GLM-5.30.0000$17.6732.5M199k19h 29m2152
87Muse Spark 1.30.0000$12.9112.2M103k19h 52m1421
88DeepSeek V4 Flash Vision Exp0.0000$0.7637.0M151k19h 59m2751
89Gemini 3.8 Flash0.0000$11.9258.4M116k9h 56m3017
90Qwen3.8-Max0.0000$7.4421.5M284k19h 29m2532
91Grok 4.60.0000$18.5324.3M327k17h 14m1947
92DeepSeek V4 Flash Vision Exp0.0000$1.2360.8M248k19h 49m3231
93Qwen3.8-Max-09020.0000$21.7360.3M968k19h 32m3663
94Grok 4.60.0000$32.8828.8M249k18h 12m1965
95Muse Spark 1.30.0000$30.0827.4M216k19h 49m2041
96Gemini 3.7 Flash0.0000$8.3027.1M102k9h 38m18810
97GPT-6 Astra0.0000$142.9574.8M253k19h 46m3117
98Grok 4.60.0000$23.9518.0M271k19h 15m1454
99Inkling0.0000$4.164.3M15k34m1073
100Kimi K30.0000$10.948.7M101k18h 40m1211
101Gemini 3.8 Flash0.0000$9.4651.9M115k6h 1m2686
102Claude Sonnet 5.50.0000$32.34129.9M492k18h 23m3262
103Grok 4.70.0000$29.8818.4M159k14h 59m1193
104Gemini 3.7 Flash0.0000$30.3173.0M132k18h 49m31013
105GLM-5.3 Flash0.0000$2.3142.4M198k17h 43m2612
106Qwen3.8-Max-09020.0000$37.59115.3M1.0M19h 35m5413
107Grok 4.70.0000$36.1819.8M164k20h 0m1265
108Gemini 3.8 Flash0.0000$14.9862.0M118k8h 40m2957
109Grok 4.70.0000$59.6535.5M185k17h 44m1703
110Grok 4.60.0000$20.0718.4M217k18h 24m1568