GPT-6 Astra在驾驶基准测试中表现领先
GPT-6 Astra has gained the ability to drive a car
多模型在具身智能/驾驶场景的横向实测数据，GPT-6 Astra 首次跑通全程，值得 Agent 研究者关注其工具调用与规划能力的跃迁。
Leaderboard
排行榜
Up to 3 attempts in one continuous chat. Expand a model to inspect its runs, then click a run to view its trace and video.
在同一个连续聊天中最多尝试 3 次。展开模型以检查其运行记录，然后单击某次运行以查看其轨迹和视频。
rank bybest of up to 3 (same context)first attempt
按最多 3 次中的最佳成绩排名（相同上下文）首次尝试
ModelBest progress
模型最佳进展
Finish time
完成时间
Attempt 1Attempt 2Attempt 3
第 1 次尝试第 2 次尝试第 3 次尝试
GPT-6 Astra
Codex · medium
100%
in5:22
耗时 5:22
49%100%5:22—
ExpandCollapse
展开折叠
attemptprogressHow far along the course centerline the attempt got while staying within 4 m of it, as a share of the centerline length to the finish zone; a collision keeps the progress reached before it.distanceGPS speed integrated from the first accepted set_motion to the end of the attempt's last engagement.finish timeFirst accepted set_motion to the end of the attempt's last engagement. Shown for completed runs only; DNF means did not finish.commandsAccepted set_motion and stop_now calls during the attempt.tokens · costTotal tokens and cost at list prices for the attempt, including the reflection after it.
attemptprogress：尝试过程中保持在赛道中心线 4 米范围内的距离占至终点区中心线长度的比例；发生碰撞时保留碰撞前达到的进度。distance：从第一次被接受的 set_motion 到尝试最后一次交互结束期间集成的 GPS 速度。finish time：从第一次被接受的 set_motion 到尝试最后一次交互结束的时间。仅显示已完成运行的数据；DNF 表示未完成。commands：尝试期间接受的 set_motion 和 stop_now 调用次数。tokens · cost：尝试的总 token 数和按列表价格计算的成本，包括尝试后的反思部分。
same chat
同一聊天
#1
49%
67.3 mDNF81.2M · $2.01
67.3 米 DNF 81.2M · $2.01
#2
100%
134.7 m5:22246.6M · $7.74
134.7 米 5:22 246.6M · $7.74
Claude Fable 5.1
Claude Code · medium
45%
9%10%45%
ExpandCollapse
展开折叠
attemptprogressHow far along the course centerline the attempt got while staying within 4 m of it, as a share of the centerline length to the finish zone; a collision keeps the progress reached before it.distanceGPS speed integrated from the first accepted set_motion to the end of the attempt's last engagement.finish timeFirst accepted set_motion to the end of the attempt's last engagement. Shown for completed runs only; DNF means did not finish.commandsAccepted set_motion and stop_now calls during the attempt.tokens · costTotal tokens and cost at list prices for the attempt, including the reflection after it.
attemptprogress：尝试过程中保持在赛道中心线 4 米范围内的距离占至终点区中心线长度的比例；发生碰撞时保留碰撞前达到的进度。distance：从第一次被接受的 set_motion 到尝试最后一次交互结束期间集成的 GPS 速度。finish time：从第一次被接受的 set_motion 到尝试最后一次交互结束的时间。仅显示已完成运行的数据；DNF 表示未完成。commands：尝试期间接受的 set_motion 和 stop_now 调用次数。tokens · cost：尝试的总 token 数和按列表价格计算的成本，包括尝试后的反思部分。
same chat
同一聊天
#1
9%
17.5 mDNF3581k · $0.96
17.5 米 DNF 3581k · $0.96
#2
10%
27.3 mDNF4866k · $1.35
27.3 米 DNF 4866k · $1.35
#3
45%
73.7 mDNF82.1M · $1.64
73.7 米 DNF 82.1M · $1.64
Grok 4.6
Cursor · medium
11%
8%11%10%
ExpandCollapse
展开/折叠
attemptprogressHow far along the course centerline the attempt got while staying within 4 m of it, as a share of the centerline length to the finish zone; a collision keeps the progress reached before it.distanceGPS speed integrated from the first accepted set_motion to the end of the attempt's last engagement.finish timeFirst accepted set_motion to the end of the attempt's last engagement. Shown for completed runs only; DNF means did not finish.commandsAccepted set_motion and stop_now calls during the attempt.tokens · costTotal tokens and cost at list prices for the attempt, including the reflection after it.
attemptprogress：尝试过程中在距离赛道中心线4米范围内行进的距离占至终点区中心线长度的比例；发生碰撞时，进度将保留在碰撞前的数值。distance：从第一次被接受的 set_motion 到尝试最后一次交互结束期间积分的GPS速度。finish time：从第一次被接受的 set_motion 到尝试最后一次交互结束的时间段。仅显示已完成运行的数据；DNF 表示未完成（did not finish）。commands：尝试期间接受的 set_motion 和 stop_now 调用。tokens · cost：尝试的总 token 数和按列表价格计算的成本，包括事后的退款。
same chat
同一聊天
#1
8%
14.4 mDNF2216k · $0.18
14.4 m DNF 2216k · $0.18
#2
11%
22.6 mDNF3294k · $0.19
22.6 m DNF 3294k · $0.19
#3
10%
22.2 mDNF3516k · $0.29
22.2 m DNF 3516k · $0.29
GPT-5.6 Sol
Codex · medium
6%
6%6%6%
ExpandCollapse
展开/折叠
attemptprogressHow far along the course centerline the attempt got while staying within 4 m of it, as a share of the centerline length to the finish zone; a collision keeps the progress reached before it.distanceGPS speed integrated from the first accepted set_motion to the end of the attempt's last engagement.finish timeFirst accepted set_motion to the end of the attempt's last engagement. Shown for completed runs only; DNF means did not finish.commandsAccepted set_motion and stop_now calls during the attempt.tokens · costTotal tokens and cost at list prices for the attempt, including the reflection after it.
attemptprogress：尝试过程中在距离赛道中心线4米范围内行进的距离占至终点区中心线长度的比例；发生碰撞时，进度将保留在碰撞前的数值。distance：从第一次被接受的 set_motion 到尝试最后一次交互结束期间积分的GPS速度。finish time：从第一次被接受的 set_motion 到尝试最后一次交互结束的时间段。仅显示已完成运行的数据；DNF 表示未完成（did not finish）。commands：尝试期间接受的 set_motion 和 stop_now 调用。tokens · cost：尝试的总 token 数和按列表价格计算的成本，包括事后的退款。
same chat
同一聊天
#1
6%
15.0 mDNF3364k · $0.35
15.0 m DNF 3364k · $0.35
#2
6%
15.6 mDNF4729k · $0.43
15.6 m DNF 4729k · $0.43
#3
6%
17.1 mDNF2517k · $0.27
17.1 m DNF 2517k · $0.27
Explore the trackView the trajectory replays↓
探索赛道查看轨迹回放↓
Explore the course
探索赛道
更进一步：量化金融体系
看懂新闻只是起点——沿量化金融路径，把它变成能交付的工程能力