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-5.6-terra

Model reported by the benchmark

Overall score
74.99 分
Long tasks score
48.4 分
Inference score
90.6 分
Coding score
86.2 分
More metrics
Browser task score
84 分
Data analysis score
79.3 分
Research score
60.4 分
Documents & design score
76 分
Overall cost per task
1.268075 USD / Task

Total cost $601.0676 · 474 tasks

Long tasks cost per task
11.756906 USD / Task

Total cost $211.6243 · 18 tasks

Inference cost per task
0.233939 USD / Task

Total cost $47.2556 · 202 tasks

Coding cost per task
2.714731 USD / Task

Total cost $78.7272 · 29 tasks

Browsing cost per task
0.801192 USD / Task

Total cost $20.0298 · 25 tasks

Data analysis cost per task
0.277761 USD / Task

Total cost $41.6642 · 150 tasks

Research cost per task
7.439208 USD / Task

Total cost $185.9802 · 25 tasks

Documents & design cost per task
0.631452 USD / Task

Total cost $15.7863 · 25 tasks

gpt-6.1-sol

Model reported by the benchmark

Overall score
85.41 分
Long tasks score
73 分
Inference score
92.6 分
Coding score
96 分
More metrics
Browser task score
88 分
Data analysis score
82.7 分
Research score
83.5 分
Documents & design score
82.1 分
Overall cost per task
0.627097 USD / Task

Total cost $297.2439 · 474 tasks

Long tasks cost per task
6.487661 USD / Task

Total cost $116.7779 · 18 tasks

Inference cost per task
0.080313 USD / Task

Total cost $16.2233 · 202 tasks

Coding cost per task
0.482069 USD / Task

Total cost $13.98 · 29 tasks

Browsing cost per task
2.4608 USD / Task

Total cost $61.52 · 25 tasks

Data analysis cost per task
0.145418 USD / Task

Total cost $21.8127 · 150 tasks

Research cost per task
2.3068 USD / Task

Total cost $57.67 · 25 tasks

Documents & design cost per task
0.3704 USD / Task

Total cost $9.26 · 25 tasks

gpt-5.6-luna

Model reported by the benchmark

Overall score
66.09 分
Long tasks score
30.4 分
Inference score
85.6 分
Coding score
73.6 分
More metrics
Browser task score
68 分
Data analysis score
78 分
Research score
51 分
Documents & design score
76 分
Overall cost per task
0.457984 USD / Task

Total cost $217.0843 · 474 tasks

Long tasks cost per task
1.101161 USD / Task

Total cost $19.8209 · 18 tasks

Inference cost per task
0.138451 USD / Task

Total cost $27.9672 · 202 tasks

Coding cost per task
1.823203 USD / Task

Total cost $52.8729 · 29 tasks

Browsing cost per task
0.370004 USD / Task

Total cost $9.2501 · 25 tasks

Data analysis cost per task
0.133729 USD / Task

Total cost $20.0593 · 150 tasks

Research cost per task
3.130584 USD / Task

Total cost $78.2646 · 25 tasks

Documents & design cost per task
0.353972 USD / Task

Total cost $8.8493 · 25 tasks

deepseek-v4.1-flash

Model reported by the benchmark

Overall score
78.04 分
Long tasks score
40.1 分
Inference score
86.7 分
Coding score
90.2 分
More metrics
Browser task score
84 分
Data analysis score
79.3 分
Research score
86.8 分
Documents & design score
79.2 分
Overall cost per task
0.083523 USD / Task

Total cost $39.59 · 474 tasks

Long tasks cost per task
0.822417 USD / Task

Total cost $14.8035 · 18 tasks

Inference cost per task
0.022826 USD / Task

Total cost $4.6108 · 202 tasks

Coding cost per task
0.118859 USD / Task

Total cost $3.4469 · 29 tasks

Browsing cost per task
0.174324 USD / Task

Total cost $4.3581 · 25 tasks

Data analysis cost per task
0.035553 USD / Task

Total cost $5.333 · 150 tasks

Research cost per task
0.194104 USD / Task

Total cost $4.8526 · 25 tasks

Documents & design cost per task
0.087404 USD / Task

Total cost $2.1851 · 25 tasks

gemma-4-31b

Model reported by the benchmark

Overall score
41.53 分
Long tasks score
13.3 分
Inference score
68.3 分
Coding score
49 分
More metrics
Browser task score
4 分
Data analysis score
53.3 分
Research score
44.4 分
Documents & design score
58.4 分
Overall cost per task
0.071171 USD / Task

Total cost $33.735 · 474 tasks

Long tasks cost per task
0.988689 USD / Task

Total cost $17.7964 · 18 tasks

Inference cost per task
0.00377 USD / Task

Total cost $0.7615 · 202 tasks

Coding cost per task
0.23261 USD / Task

Total cost $6.7457 · 29 tasks

Browsing cost per task
0.25412 USD / Task

Total cost $6.353 · 25 tasks

Data analysis cost per task
0.004565 USD / Task

Total cost $0.6847 · 150 tasks

Research cost per task
0.040956 USD / Task

Total cost $1.0239 · 25 tasks

Documents & design cost per task
0.014792 USD / Task

Total cost $0.3698 · 25 tasks

claude-opus-5

Model reported by the benchmark

Overall score
82.09 分
Long tasks score
62.1 分
Inference score
91.2 分
Coding score
90.8 分
More metrics
Browser task score
84 分
Data analysis score
74.6 分
Research score
88.9 分
Documents & design score
83 分
Overall cost per task
3.207076 USD / Task

Total cost $1520.1541 · 474 tasks

Long tasks cost per task
48.753172 USD / Task

Total cost $877.5571 · 18 tasks

Inference cost per task
0.168446 USD / Task

Total cost $34.026 · 202 tasks

Coding cost per task
3.921766 USD / Task

Total cost $113.7312 · 29 tasks

Browsing cost per task
2.965252 USD / Task

Total cost $74.1313 · 25 tasks

Data analysis cost per task
0.745529 USD / Task

Total cost $111.8294 · 150 tasks

Research cost per task
8.900476 USD / Task

Total cost $222.5119 · 25 tasks

Documents & design cost per task
3.454688 USD / Task

Total cost $86.3672 · 25 tasks

claude-sonnet-4.6

Model reported by the benchmark

Overall score
65.09 分
Long tasks score
23.1 分
Inference score
86.4 分
Coding score
66.6 分
More metrics
Browser task score
68 分
Data analysis score
76.1 分
Research score
63.6 分
Documents & design score
71.8 分
Overall cost per task
Not provided
Long tasks cost per task
20.419611 USD / Task

Total cost $367.553 · 18 tasks

Inference cost per task
Not provided
Coding cost per task
2.628662 USD / Task

Total cost $76.2312 · 29 tasks

Browsing cost per task
1.856448 USD / Task

Total cost $46.4112 · 25 tasks

Data analysis cost per task
Not provided
Research cost per task
2.090976 USD / Task

Total cost $52.2744 · 25 tasks

Documents & design cost per task
1.4658 USD / Task

Total cost $36.645 · 25 tasks

gemini-4-argon

Model reported by the benchmark

Overall score
84.1 分
Long tasks score
69.9 分
Inference score
95.1 分
Coding score
89.7 分
More metrics
Browser task score
92 分
Data analysis score
85.2 分
Research score
83.1 分
Documents & design score
73.7 分
Overall cost per task
1.5394 USD / Task

Total cost $729.675699 · 474 tasks

Long tasks cost per task
13.841872 USD / Task

Total cost $249.1537 · 18 tasks

Inference cost per task
0.093203 USD / Task

Total cost $18.827034 · 202 tasks

Coding cost per task
5.076328 USD / Task

Total cost $147.2135 · 29 tasks

Browsing cost per task
3.433868 USD / Task

Total cost $85.8467 · 25 tasks

Data analysis cost per task
0.338232 USD / Task

Total cost $50.734865 · 150 tasks

Research cost per task
5.955556 USD / Task

Total cost $148.8889 · 25 tasks

Documents & design cost per task
1.16044 USD / Task

Total cost $29.011 · 25 tasks

gpt-5.5

Model reported by the benchmark

Overall score
68.7 分
Long tasks score
44.6 分
Inference score
89.7 分
Coding score
85.1 分
More metrics
Browser task score
48 分
Data analysis score
81.6 分
Research score
61.6 分
Documents & design score
70.3 分
Overall cost per task
2.078327 USD / Task

Total cost $985.1271 · 474 tasks

Long tasks cost per task
13.851483 USD / Task

Total cost $249.3267 · 18 tasks

Inference cost per task
0.289229 USD / Task

Total cost $58.4242 · 202 tasks

Coding cost per task
4.207645 USD / Task

Total cost $122.0217 · 29 tasks

Browsing cost per task
4.819536 USD / Task

Total cost $120.4884 · 25 tasks

Data analysis cost per task
0.4055 USD / Task

Total cost $60.825 · 150 tasks

Research cost per task
13.950244 USD / Task

Total cost $348.7561 · 25 tasks

Documents & design cost per task
1.0114 USD / Task

Total cost $25.285 · 25 tasks