Find the right AI for you

Search by name or tell us what you want to do
API servicesChoose how to access modelsAI modelsCapabilities, specifications, and pricingAI toolsFind useful tools by taskMCP directoryConnect tools, data, and workflowsAgent skillsFind reusable skills for AI agentsBenchmarksCompare model performanceOnline toolsConvert accounts, format JSON, estimate costs and test APIsModel timelineExplore model releasesUse casesStart with a practical task
EnterSearch · ↑↓Choose a resultEscClose
Submit中文
Project benchmark

LiveBench 2.0 task results & run costs

LiveBench · 49 configurations with published scores. Review each metric and its test conditions separately.

Report identity & conditions

Data as of2026-10-18· Report published—. See individual records for run dates and test conditions.

Leaderboard updated

  • Each run retains a 0–100 score for each of seven task categories. The original report assigns zero to execution failures that remain unresolved after retries. Missing cells remain missing; we do not fill them with zero.
  • The overall score is the equal-weight mean of seven task categories. Our comparison includes only configurations with all seven scores, keeping partial averages separate.
  • Costs cover all calls to the tested model at provider list prices, including cached input rates. Fixed initial state and supporting framework calls are excluded. Each task retains its total cost and count: USD/task = total cost ÷ task count. This is neither cost per successful task nor a current API unit price.
  • Overall cost per task is the sum of costs divided by the sum of task counts across all seven categories. It is provided only when every category has complete costs and samples, avoiding direct comparisons between different task mixes.
  • Models are shown under the report’s configuration identifiers. Shared names or mutable API aliases do not prove an exact snapshot. Test dates, parameters and settings follow each run’s documentation.
  • Results come from the official LiveBench leaderboard. Follow the report link for methodology and original versions. We do not reproduce test questions, answers, benchmark code or API prices without a defined scope.
  • Benchmark version: 2026-09-02; leaderboard updated: 2026-10-14. Neither is an individual test date or an official model release date.
Read the original report ↗

Compare metrics within the same report. Each chart explains its direction and scope.
See the details for full values and run conditions.

Overall score

Equal-weight mean across seven task categories. Only configurations with all seven scores enter this overall comparison.: showing 8 of 49 records with values. See details for configurations and dates.

Chart data · 8 entries
Name points
claude-fable-5.187.99
claude-opus-5.586.53
gpt-6-astra86.14
gpt-6.1-sol85.41
gemini-4-argon84.1
claude-fable-583.61
claude-sonnet-5.582.46
claude-opus-582.09

Long tasks score

The run’s mean score for this task, from 0 to 100. It is not ranked directly against other benchmarks or tasks.: showing 8 of 49 records with values. See details for configurations and dates.

Chart data · 8 entries
Name points
claude-opus-5.585
gpt-6-astra82.4
claude-fable-5.180.8
claude-sonnet-5.578.8
gpt-6.1-sol73
gemini-4-argon69.9
gpt-6-sol66.9
claude-opus-562.1

Full results & model versions

49 / 49configurations have published scores in this scope. Scored configurations are listed; missing values remain “Not provided.” Different run configurations of the same model are listed separately.

gpt-6-astra

Model reported by the benchmark

Overall score
86.14 分
Long tasks score
82.4 分
Inference score
92.7 分
Coding score
93.7 分
More metrics
Browser task score
84 分
Data analysis score
83 分
Research score
85.2 分
Documents & design score
82 分
Overall cost per task
2.964769 USD / Task

Total cost $1402.3356 · 473 tasks

Long tasks cost per task
38.378918 USD / Task

Total cost $652.4416 · 17 tasks

Inference cost per task
0.389843 USD / Task

Total cost $78.7482 · 202 tasks

Coding cost per task
3.315369 USD / Task

Total cost $96.1457 · 29 tasks

Browsing cost per task
5.341308 USD / Task

Total cost $133.5327 · 25 tasks

Data analysis cost per task
0.698902 USD / Task

Total cost $104.8353 · 150 tasks

Research cost per task
11.126884 USD / Task

Total cost $278.1721 · 25 tasks

Documents & design cost per task
2.3384 USD / Task

Total cost $58.46 · 25 tasks

qwen3.8-27b

Model reported by the benchmark

Overall score
62.66 分
Long tasks score
27.9 分
Inference score
80 分
Coding score
86.8 分
More metrics
Browser task score
32 分
Data analysis score
76.6 分
Research score
59.4 分
Documents & design score
75.9 分
Overall cost per task
0.415958 USD / Task

Total cost $197.1643 · 474 tasks

Long tasks cost per task
6.256917 USD / Task

Total cost $112.6245 · 18 tasks

Inference cost per task
0.016407 USD / Task

Total cost $3.3142 · 202 tasks

Coding cost per task
0.604559 USD / Task

Total cost $17.5322 · 29 tasks

Browsing cost per task
1.10966 USD / Task

Total cost $27.7415 · 25 tasks

Data analysis cost per task
0.019885 USD / Task

Total cost $2.9827 · 150 tasks

Research cost per task
0.964912 USD / Task

Total cost $24.1228 · 25 tasks

Documents & design cost per task
0.353856 USD / Task

Total cost $8.8464 · 25 tasks

beam-1.0

Model reported by the benchmark

Overall score
55.5 分
Long tasks score
17.3 分
Inference score
80.2 分
Coding score
83.9 分
More metrics
Browser task score
0 分
Data analysis score
66.4 分
Research score
72.8 分
Documents & design score
67.9 分
Overall cost per task
Not provided
Long tasks cost per task
Not provided
Inference cost per task
Not provided
Coding cost per task
Not provided
Browsing cost per task
Not provided
Data analysis cost per task
Not provided
Research cost per task
Not provided
Documents & design cost per task
Not provided

claude-haiku-4.5

Model reported by the benchmark

Overall score
46.04 分
Long tasks score
11.1 分
Inference score
61.7 分
Coding score
48.1 分
More metrics
Browser task score
40 分
Data analysis score
59.3 分
Research score
42.2 分
Documents & design score
59.9 分
Overall cost per task
Not provided
Long tasks cost per task
2.548517 USD / Task

Total cost $45.8733 · 18 tasks

Inference cost per task
Not provided
Coding cost per task
0.494024 USD / Task

Total cost $14.3267 · 29 tasks

Browsing cost per task
0.812664 USD / Task

Total cost $20.3166 · 25 tasks

Data analysis cost per task
Not provided
Research cost per task
0.257204 USD / Task

Total cost $6.4301 · 25 tasks

Documents & design cost per task
0.106652 USD / Task

Total cost $2.6663 · 25 tasks

gpt-6-sol

Model reported by the benchmark

Overall score
81.91 分
Long tasks score
66.9 分
Inference score
88.7 分
Coding score
93.1 分
More metrics
Browser task score
92 分
Data analysis score
81.2 分
Research score
69.3 分
Documents & design score
82.2 分
Overall cost per task
0.899811 USD / Task

Total cost $426.5103 · 474 tasks

Long tasks cost per task
8.621461 USD / Task

Total cost $155.1863 · 18 tasks

Inference cost per task
0.120914 USD / Task

Total cost $24.4247 · 202 tasks

Coding cost per task
0.976897 USD / Task

Total cost $28.33 · 29 tasks

Browsing cost per task
3.806 USD / Task

Total cost $95.15 · 25 tasks

Data analysis cost per task
0.198929 USD / Task

Total cost $29.8393 · 150 tasks

Research cost per task
3.318 USD / Task

Total cost $82.95 · 25 tasks

Documents & design cost per task
0.4252 USD / Task

Total cost $10.63 · 25 tasks

claude-opus-5.5

Model reported by the benchmark

Overall score
86.53 分
Long tasks score
85 分
Inference score
92.2 分
Coding score
96.3 分
More metrics
Browser task score
80 分
Data analysis score
80.3 分
Research score
85 分
Documents & design score
86.9 分
Overall cost per task
3.647876 USD / Task

Total cost $1729.0933 · 474 tasks

Long tasks cost per task
38.008772 USD / Task

Total cost $684.1579 · 18 tasks

Inference cost per task
0.252235 USD / Task

Total cost $50.9515 · 202 tasks

Coding cost per task
5.03069 USD / Task

Total cost $145.89 · 29 tasks

Browsing cost per task
14.6696 USD / Task

Total cost $366.74 · 25 tasks

Data analysis cost per task
0.814559 USD / Task

Total cost $122.1839 · 150 tasks

Research cost per task
8.6184 USD / Task

Total cost $215.46 · 25 tasks

Documents & design cost per task
5.7484 USD / Task

Total cost $143.71 · 25 tasks

claude-fable-5.1

Model reported by the benchmark

Overall score
87.99 分
Long tasks score
80.8 分
Inference score
91.7 分
Coding score
97.7 分
More metrics
Browser task score
100 分
Data analysis score
80.3 分
Research score
82 分
Documents & design score
83.4 分
Overall cost per task
3.982806 USD / Task

Total cost $1887.85 · 474 tasks

Long tasks cost per task
51.572789 USD / Task

Total cost $928.3102 · 18 tasks

Inference cost per task
0.460136 USD / Task

Total cost $92.9474 · 202 tasks

Coding cost per task
5.282366 USD / Task

Total cost $153.1886 · 29 tasks

Browsing cost per task
3.746228 USD / Task

Total cost $93.6557 · 25 tasks

Data analysis cost per task
1.333143 USD / Task

Total cost $199.9714 · 150 tasks

Research cost per task
11.613816 USD / Task

Total cost $290.3454 · 25 tasks

Documents & design cost per task
5.177252 USD / Task

Total cost $129.4313 · 25 tasks

kimi-k3

Model reported by the benchmark

Overall score
63.39 分
Long tasks score
42.8 分
Inference score
90.7 分
Coding score
59.3 分
More metrics
Browser task score
16 分
Data analysis score
78.7 分
Research score
79 分
Documents & design score
77.2 分
Overall cost per task
1.164515 USD / Task

Total cost $551.98 · 474 tasks

Long tasks cost per task
17.990494 USD / Task

Total cost $323.8289 · 18 tasks

Inference cost per task
0.195096 USD / Task

Total cost $39.4094 · 202 tasks

Coding cost per task
2.061386 USD / Task

Total cost $59.7802 · 29 tasks

Browsing cost per task
0.382016 USD / Task

Total cost $9.5504 · 25 tasks

Data analysis cost per task
0.352355 USD / Task

Total cost $52.8533 · 150 tasks

Research cost per task
1.763404 USD / Task

Total cost $44.0851 · 25 tasks

Documents & design cost per task
0.898908 USD / Task

Total cost $22.4727 · 25 tasks

muse-spark-open

Model reported by the benchmark

Overall score
57.57 分
Long tasks score
18.9 分
Inference score
73.7 分
Coding score
73.6 分
More metrics
Browser task score
36 分
Data analysis score
60.2 分
Research score
80.3 分
Documents & design score
60.3 分
Overall cost per task
0.488394 USD / Task

Total cost $231.498567 · 474 tasks

Long tasks cost per task
4.704306 USD / Task

Total cost $84.6775 · 18 tasks

Inference cost per task
0.050127 USD / Task

Total cost $10.125671 · 202 tasks

Coding cost per task
1.252538 USD / Task

Total cost $36.3236 · 29 tasks

Browsing cost per task
2.643368 USD / Task

Total cost $66.0842 · 25 tasks

Data analysis cost per task
0.048547 USD / Task

Total cost $7.282096 · 150 tasks

Research cost per task
0.724588 USD / Task

Total cost $18.1147 · 25 tasks

Documents & design cost per task
0.355632 USD / Task

Total cost $8.8908 · 25 tasks

glm-5.3

Model reported by the benchmark

Overall score
60.74 分
Long tasks score
35.9 分
Inference score
85.8 分
Coding score
82.8 分
More metrics
Browser task score
0 分
Data analysis score
70.2 分
Research score
73.2 分
Documents & design score
77.3 分
Overall cost per task
Not provided
Long tasks cost per task
18.381939 USD / Task

Total cost $330.8749 · 18 tasks

Inference cost per task
0.264766 USD / Task

Total cost $53.4827 · 202 tasks

Coding cost per task
2.138393 USD / Task

Total cost $62.0134 · 29 tasks

Browsing cost per task
Not provided
Data analysis cost per task
0.373101 USD / Task

Total cost $55.9652 · 150 tasks

Research cost per task
1.634592 USD / Task

Total cost $40.8648 · 25 tasks

Documents & design cost per task
0.820736 USD / Task

Total cost $20.5184 · 25 tasks

gpt-6-luna

Model reported by the benchmark

Overall score
69.4 分
Long tasks score
48.1 分
Inference score
81.8 分
Coding score
84.5 分
More metrics
Browser task score
72 分
Data analysis score
73.4 分
Research score
43.5 分
Documents & design score
82.5 分
Overall cost per task
0.091607 USD / Task

Total cost $43.4217 · 474 tasks

Long tasks cost per task
0.967611 USD / Task

Total cost $17.417 · 18 tasks

Inference cost per task
0.021405 USD / Task

Total cost $4.3238 · 202 tasks

Coding cost per task
0.095172 USD / Task

Total cost $2.76 · 29 tasks

Browsing cost per task
0.2816 USD / Task

Total cost $7.04 · 25 tasks

Data analysis cost per task
0.016606 USD / Task

Total cost $2.4909 · 150 tasks

Research cost per task
0.3452 USD / Task

Total cost $8.63 · 25 tasks

Documents & design cost per task
0.0304 USD / Task

Total cost $0.76 · 25 tasks

smaug-flash

Model reported by the benchmark

Overall score
57.54 分
Long tasks score
27.1 分
Inference score
86.2 分
Coding score
60.1 分
More metrics
Browser task score
0 分
Data analysis score
79.1 分
Research score
74.1 分
Documents & design score
76.2 分
Overall cost per task
Not provided
Long tasks cost per task
Not provided
Inference cost per task
0.008641 USD / Task

Total cost $1.7455 · 202 tasks

Coding cost per task
0.17231 USD / Task

Total cost $4.997 · 29 tasks

Browsing cost per task
Not provided
Data analysis cost per task
0.008367 USD / Task

Total cost $1.255 · 150 tasks

Research cost per task
0.127652 USD / Task

Total cost $3.1913 · 25 tasks

Documents & design cost per task
0.032836 USD / Task

Total cost $0.8209 · 25 tasks

muse-spark-1.3

Model reported by the benchmark

Overall score
74.76 分
Long tasks score
40.4 分
Inference score
89.7 分
Coding score
90.2 分
More metrics
Browser task score
80 分
Data analysis score
79.6 分
Research score
68.6 分
Documents & design score
74.8 分
Overall cost per task
0.747174 USD / Task

Total cost $354.1606 · 474 tasks

Long tasks cost per task
8.83895 USD / Task

Total cost $159.1011 · 18 tasks

Inference cost per task
0.135722 USD / Task

Total cost $27.4159 · 202 tasks

Coding cost per task
1.371617 USD / Task

Total cost $39.7769 · 29 tasks

Browsing cost per task
1.401388 USD / Task

Total cost $35.0347 · 25 tasks

Data analysis cost per task
0.169429 USD / Task

Total cost $25.4143 · 150 tasks

Research cost per task
2.071076 USD / Task

Total cost $51.7769 · 25 tasks

Documents & design cost per task
0.625632 USD / Task

Total cost $15.6408 · 25 tasks

claude-opus-4.8

Model reported by the benchmark

Overall score
76.77 分
Long tasks score
48.3 分
Inference score
89.2 分
Coding score
77.3 分
More metrics
Browser task score
96 分
Data analysis score
66 分
Research score
81.8 分
Documents & design score
78.8 分
Overall cost per task
3.38177 USD / Task

Total cost $1602.959 · 474 tasks

Long tasks cost per task
51.969644 USD / Task

Total cost $935.4536 · 18 tasks

Inference cost per task
0.605731 USD / Task

Total cost $122.3576 · 202 tasks

Coding cost per task
4.151659 USD / Task

Total cost $120.3981 · 29 tasks

Browsing cost per task
5.10332 USD / Task

Total cost $127.583 · 25 tasks

Data analysis cost per task
0.645581 USD / Task

Total cost $96.8372 · 150 tasks

Research cost per task
3.764684 USD / Task

Total cost $94.1171 · 25 tasks

Documents & design cost per task
4.248496 USD / Task

Total cost $106.2124 · 25 tasks

muse-wonder-1.0

Model reported by the benchmark

Overall score
78.1 分
Long tasks score
53.2 分
Inference score
90.6 分
Coding score
86.2 分
More metrics
Browser task score
72 分
Data analysis score
81.6 分
Research score
88 分
Documents & design score
75.1 分
Overall cost per task
1.411194 USD / Task

Total cost $668.9058 · 474 tasks

Long tasks cost per task
14.482883 USD / Task

Total cost $260.6919 · 18 tasks

Inference cost per task
0.298363 USD / Task

Total cost $60.2693 · 202 tasks

Coding cost per task
3.102252 USD / Task

Total cost $89.9653 · 29 tasks

Browsing cost per task
4.138132 USD / Task

Total cost $103.4533 · 25 tasks

Data analysis cost per task
0.365737 USD / Task

Total cost $54.8605 · 150 tasks

Research cost per task
3.299512 USD / Task

Total cost $82.4878 · 25 tasks

Documents & design cost per task
0.687108 USD / Task

Total cost $17.1777 · 25 tasks

qwen3.8-max

Model reported by the benchmark

Overall score
64.24 分
Long tasks score
42.3 分
Inference score
88.2 分
Coding score
60.3 分
More metrics
Browser task score
32 分
Data analysis score
78.4 分
Research score
69.9 分
Documents & design score
78.6 分
Overall cost per task
1.000063 USD / Task

Total cost $474.03 · 474 tasks

Long tasks cost per task
14.93005 USD / Task

Total cost $268.7409 · 18 tasks

Inference cost per task
0.131491 USD / Task

Total cost $26.5611 · 202 tasks

Coding cost per task
1.986917 USD / Task

Total cost $57.6206 · 29 tasks

Browsing cost per task
1.411484 USD / Task

Total cost $35.2871 · 25 tasks

Data analysis cost per task
0.218945 USD / Task

Total cost $32.8418 · 150 tasks

Research cost per task
1.586148 USD / Task

Total cost $39.6537 · 25 tasks

Documents & design cost per task
0.532992 USD / Task

Total cost $13.3248 · 25 tasks

muse-spark-1.2

Model reported by the benchmark

Overall score
62.4 分
Long tasks score
21.3 分
Inference score
90 分
Coding score
78.2 分
More metrics
Browser task score
36 分
Data analysis score
76.5 分
Research score
67.7 分
Documents & design score
67.1 分
Overall cost per task
0.536117 USD / Task

Total cost $254.1195 · 474 tasks

Long tasks cost per task
3.948783 USD / Task

Total cost $71.0781 · 18 tasks

Inference cost per task
0.099368 USD / Task

Total cost $20.0724 · 202 tasks

Coding cost per task
1.465721 USD / Task

Total cost $42.5059 · 29 tasks

Browsing cost per task
2.049188 USD / Task

Total cost $51.2297 · 25 tasks

Data analysis cost per task
0.109447 USD / Task

Total cost $16.4171 · 150 tasks

Research cost per task
1.448716 USD / Task

Total cost $36.2179 · 25 tasks

Documents & design cost per task
0.663936 USD / Task

Total cost $16.5984 · 25 tasks

deepseek-v4-flash-vision

Model reported by the benchmark

Overall score
63.63 分
Long tasks score
39 分
Inference score
85.4 分
Coding score
37.6 分
More metrics
Browser task score
64 分
Data analysis score
79.5 分
Research score
69.2 分
Documents & design score
70.7 分
Overall cost per task
0.085098 USD / Task

Total cost $40.3366 · 474 tasks

Long tasks cost per task
1.040689 USD / Task

Total cost $18.7324 · 18 tasks

Inference cost per task
0.042434 USD / Task

Total cost $8.5717 · 202 tasks

Coding cost per task
0.040134 USD / Task

Total cost $1.1639 · 29 tasks

Browsing cost per task
0.13204 USD / Task

Total cost $3.301 · 25 tasks

Data analysis cost per task
0.039525 USD / Task

Total cost $5.9288 · 150 tasks

Research cost per task
0.083132 USD / Task

Total cost $2.0783 · 25 tasks

Documents & design cost per task
0.02242 USD / Task

Total cost $0.5605 · 25 tasks

grok-4.5

Model reported by the benchmark

Overall score
70.06 分
Long tasks score
27.9 分
Inference score
87.2 分
Coding score
94.3 分
More metrics
Browser task score
44 分
Data analysis score
73 分
Research score
85.9 分
Documents & design score
78.1 分
Overall cost per task
0.518137 USD / Task

Total cost $245.5971 · 474 tasks

Long tasks cost per task
4.498294 USD / Task

Total cost $80.9693 · 18 tasks

Inference cost per task
0.087953 USD / Task

Total cost $17.7665 · 202 tasks

Coding cost per task
0.959176 USD / Task

Total cost $27.8161 · 29 tasks

Browsing cost per task
2.917696 USD / Task

Total cost $72.9424 · 25 tasks

Data analysis cost per task
0.114523 USD / Task

Total cost $17.1784 · 150 tasks

Research cost per task
0.769248 USD / Task

Total cost $19.2312 · 25 tasks

Documents & design cost per task
0.387728 USD / Task

Total cost $9.6932 · 25 tasks

qwen3.8-flash

Model reported by the benchmark

Overall score
68.79 分
Long tasks score
37.1 分
Inference score
87.4 分
Coding score
87.4 分
More metrics
Browser task score
48 分
Data analysis score
74.2 分
Research score
74.6 分
Documents & design score
72.8 分
Overall cost per task
0.135322 USD / Task

Total cost $64.1425 · 474 tasks

Long tasks cost per task
1.872044 USD / Task

Total cost $33.6968 · 18 tasks

Inference cost per task
0.022001 USD / Task

Total cost $4.4441 · 202 tasks

Coding cost per task
0.106138 USD / Task

Total cost $3.078 · 29 tasks

Browsing cost per task
0.224024 USD / Task

Total cost $5.6006 · 25 tasks

Data analysis cost per task
0.043507 USD / Task

Total cost $6.5261 · 150 tasks

Research cost per task
0.303804 USD / Task

Total cost $7.5951 · 25 tasks

Documents & design cost per task
0.128072 USD / Task

Total cost $3.2018 · 25 tasks