Which model leads livebench_agentic_coding?

claude-opus-5-max-effort leads the livebench_agentic_coding board of 46 measured models on the Vector Wire tracker at 65.2. Every score is source-attributed with a verification tier.

livebench_agentic_coding Leaderboard — Model Scores

Refresh Models —Deals —Regulatory —Sources —

Pro Mode — raw scores, one benchmark under the lensthird-party eval harness

The field · 46 measured

16.3165.2

46 models measured, most on the third-party eval harness. Claude Opus 5 Max Effort tops the board at 65.2.

46 measured·3 new this week·Lifecycle Updated Sep 2, 2026Basis

#◆Vendor$/M in$/M out

Leader — ignites whiteField — fill cools with scoreIndependently verified only46 of 46Bars show 16.31–65.2· ◆ solid = first-party or better · outlined = arm's-length

01Claude Opus 5 Max EffortAnthropic65.2$5.0/M$25.0/M$0.230

02DeepSeek V4 Flash Vision ExpDeepSeek65.1$0.44/M$1.3/M$0.014

03Qwen3.8 MaxAlibaba64.65$2.0/M$6.0/M$0.062

04Claude Fable 5 Max EffortAnthropic62.17$10.0/M$50.0/M$0.483

05Kimi K3Moonshot62.17$2.9/M$14.3/M$0.138

06Qwen3.8 Flash NextNEWAlibaba61.62$0.15/M$0.47/M$0.005

07Qwen3.8 27bAlibaba61.36$0.50/M$3.0/M$0.029

08GLM 5.3Z.ai60.91$1.4/M$4.4/M$0.048

09Claude Sonnet 5 Xhigh EffortAnthropic59.39$2.0/M$10.0/M$0.101

10Muse Spark 1.1 XhighMeta58.54$1.3/M$4.3/M$0.047

11Gemini 3.7 Flash HighGoogle58.28$0.75/M$3.8/M$0.039

12Muse Spark 1.2 XhighMeta57.58$1.3/M$4.3/M$0.048

13Grok 4.6SpaceXAI57.02$2.0/M$6.0/M$0.070

14GLM 5.3 FlashNEWZ.ai56.77$0.15/M$0.50/M$0.006

15Grok 4.5SpaceXAI56.46$2.0/M$6.0/M$0.071

16GPT 5.6 Sol MaxOpenAI56.21$4.0/M$20.0/M$0.213

17GPT 5.6 Terra MaxOpenAI54.95$2.0/M$12.0/M$0.127

18GPT 5.5 XhighOpenAI53.99$5.0/M$30.0/M$0.324

19GPT 5.4 XhighOpenAI53.84$2.5/M$15.0/M$0.163

20Ox Alpha MaxStealth52.63———

21GLM 5.2Z.ai51.77$1.4/M$4.4/M$0.056

22Claude Opus 4.7 Xhigh EffortAnthropic50.66$5.0/M$25.0/M$0.296

23Claude Opus 4.8 Max EffortAnthropic50.51$5.0/M$25.0/M$0.297

24GPT 5.2.2025.12.11 HighOpenAI50.25$1.8/M$14.0/M$0.157

25Inkling XhighThinking Machines49.39$0.95/M$4.0/M$0.051

26GPT 5.2 CodexOpenAI49.39$1.8/M$14.0/M$0.159

27Claude Opus 4.6 Thinking Auto High EffortAnthropic48.99$5.0/M$25.0/M$0.306

28Gemini 3.5 Flash HighGoogle48.99$1.5/M$9.0/M$0.107

29GPT 5.6 Luna MaxOpenAI48.43$0.20/M$1.2/M$0.014

30Kimi K2.6 ThinkingMoonshot46.92$0.97/M$4.0/M$0.053

31GPT 5.4 Nano XhighOpenAI46.77$0.20/M$1.3/M$0.016

32DeepSeek V4 Flash 0731+1 altDeepSeek46.77$0.44/M$1.3/M$0.019

33Kimi K2.7 CodeMoonshot45.66$0.66/M$3.4/M$0.044

34Gemini 3.5 Flash Lite HighGoogle45.25$0.30/M$2.5/M$0.031

35Gemini 3.1 Pro Preview HighGoogle44.14$2.0/M$12.0/M$0.159

36Qwen3.7 MaxAlibaba43.59$2.5/M$7.5/M$0.115

37Gemini 3.6 Flash HighGoogle43.43$0.75/M$3.8/M$0.052

38Claude Sonnet 4.6 Thinking Auto Medium EffortAnthropic42.63$3.0/M$15.0/M$0.211

39DeepSeek V4 ProDeepSeek42.63$1.3/M$4.0/M$0.062

40GPT 5.4 Mini XhighOpenAI41.67$0.75/M$4.5/M$0.063

41Qwen3.6 PlusAlibaba41.36$0.50/M$3.0/M$0.042

42Minimax M3MiniMax40.66$0.28/M$1.1/M$0.017

43Claude Opus 4.5.20251101 Thinking 64k High EffortAnthropic39.7$5.0/M$25.0/M$0.378

44Qwen3.6 27bAlibaba39.29$0.60/M$3.6/M$0.053

45Grok 4.3SpaceXAI18.54$1.3/M$2.5/M$0.101

46Nemotron 3 Ultra 550b A55bNEWNVIDIA16.31$0.68/M$2.7/M$0.103

◆ solid = independently verified · aggregator attested · vendor attributed — outlined = cross-reference · source-attributed. Dimmed ◆ = confidence inferred. “+N alts” = alternate disclosures in the drawer with why-demoted. Prices per million tokens; $/point = input price ÷ score. Every row opens its model dossier.