How many tracked prints does the BrowseComp family have?

BrowseComp leaderboard — the family board.

17 TRACKED VARIANTS · 1 FIRST-CLASS BOARD

One family, 17 tracked prints. Scoring configs and splits of the base board render as sections of this page, each under its own honest label.

● LIVE DATA78 MODELS ON THIS BOARDLEADER ATRIA DAWN AT 92.5

The board — BrowseCompTOP 78 OF 78 MEASURED · BEST SCORE PER MODEL, DEFAULT VARIANT

#ModelVendorScoreVLast seen

01Atria Dawnshanghai-ai-lab92.52026-09-30

02GPT-6 AstraOpenAI91.52026-09-04

03Kimi K3Moonshot91.22026-08-23

04Claude Opus 5Anthropic90.82026-08-23

05GPT-5.6 SolOpenAI90.42026-08-23

06GPT-5.5 ProOpenAI90.12026-08-23

07GPT-5.4 ProOpenAI89.32026-06-21

08Claude Opus 4.8Anthropic88.52026-08-23

09Claude Fable 5Anthropic882026-07-27

10GPT-5.6 TerraOpenAI87.52026-08-23

11Claude Mythos PreviewAnthropic86.92026-08-23

12Kimi K2.6Moonshot86.32026-08-23

13Seed 2.1 Pro PreviewByteDance86.22026-09-26

14Gemini 3.1 ProGoogle85.92026-08-24

15Claude Sonnet 5Anthropic84.72026-08-23

16GPT-5.5OpenAI84.42026-08-23

17Hy3-previewTencent84.22026-08-23

18Claude Opus 4.6Anthropic842026-08-23

19MiniMax M3MiniMax83.52026-08-23

20DeepSeek-V4-ProDeepSeek83.42026-08-23

21GPT-5.6 LunaOpenAI83.32026-08-23

22GPT-5.4OpenAI82.72026-08-23

23Claude Opus 4.7Anthropic79.82026-08-23

24GLM-5.1Z.ai79.32026-08-23

25GPT-5.2 ProOpenAI77.92026-08-23

26Inkling-SmallThinking Machines77.42026-08-23

27Seed 2.0ByteDance77.32026-09-04

28MiniMax M2.5MiniMax76.32026-08-23

29Claude Sonnet 4.6Anthropic76.22026-08-23

30GLM-5Z.ai75.92026-10-08

31Kimi K2.5Moonshot74.92026-10-08

32DeepSeek-V4-FlashDeepSeek73.22026-06-27

33Ling-3.0-flashInclusionAI72.22026-09-09

34Qwen3.5 397B A17BAlibaba692026-08-23

35Step 3.5 FlashStepFun692026-08-23

36Seed 1.8ByteDance67.62026-09-09

37GPT-5.2OpenAI65.82026-08-23

38Qwen3.5 122B A10BAlibaba63.82026-08-23

39MiniMax M2.1MiniMax622026-08-23

40Qwen3.5 35B A3BAlibaba612026-08-23

41Qwen3.5 27BAlibaba612026-08-23

42Kimi K2 ThinkingMoonshot60.22026-05-30

43Gemini 3 ProGoogle59.22026-06-15

44MiMo V2 FlashXiaomi58.32026-08-23

45GPT-5OpenAI54.92026-08-23

46MiniMax M2.7MiniMax54.12026-06-06

47Qwen3 Max (Reasoning)Alibaba53.92026-09-09

48GLM-4.7Z.ai522026-08-23

49o4-miniOpenAI51.52026-08-23

50DeepSeek-V3.2DeepSeek51.42026-10-08

51GPT-5.1OpenAI50.82026-05-18

52o3OpenAI49.72026-08-23

53Sarvam 105BSarvam AI49.52026-08-23

54Solar Pro4Upstage49.22026-08-23

55Qwen3.6 35B A3BAlibaba48.72026-08-11

56Mistral Medium 3.5 128BMistral48.62026-08-23

57GLM-4.6Z.ai45.12026-08-23

58Grok 4 FastSpaceXAI44.92026-08-23

59Nemotron 3 Ultra 550B A55BNVIDIA44.42026-09-01

60MiniMax M2MiniMax442026-08-23

61GLM-4.7-FlashZ.ai42.82026-08-23

62DeepSeek-V3.2-ExpDeepSeek40.12026-08-23

63Claude Opus 4.5Anthropic372026-06-15

64Nemotron 3.5 LightningNVIDIA37.02026-08-26

65Sarvam 30BSarvam AI35.52026-08-23

66Nemotron 3 Super 120B A12BNVIDIA31.32026-08-23

67DeepSeek-V3.1DeepSeek302026-08-23

68gpt-oss-20bOpenAI28.32026-05-16

69GLM-4.5Z.ai26.42026-08-23

70Gemma 4 26B A4BGoogle26.32026-08-11

71Claude Sonnet 4.5Anthropic24.12026-06-06

72Nemotron 3 SuperNVIDIA22.82026-08-11

73GLM-4.5-AirZ.ai21.32026-08-23

74Kimi K2 InstructMoonshot14.12026-06-06

75Nemotron 3 NanoNVIDIA13.72026-08-11

76Claude Sonnet 4Anthropic12.22026-06-06

77Gemini 2.5 ProGoogle9.92026-06-06

78Qwen3 30B A3B 2507 ThinkingAlibaba2.32026-05-16

◆ IV — INDEPENDENTLY VERIFIED · ◆ AA — AGGREGATOR · ◆ VA — VENDOR · ◇ SA — SOURCE-ATTRIBUTED

Variants of this boardSCORING CONFIGS & SPLITS — NOT CITED STANDALONE, SO THEY LIVE HERE

BrowseComp-ZH

VARIANT28 models · best score per model

#ModelVendorScoreVLast seen

01Seed 1.8ByteDance81.32026-09-09

02GPT-5.2OpenAI76.12026-05-17

03GLM-5Z.ai72.72026-05-17

04Qwen3.5 397B A17BAlibaba70.32026-08-23

05Qwen3.5 122B A10BAlibaba69.92026-08-23

06Qwen3.5 35B A3BAlibaba69.52026-08-23

07Step 3.5 FlashStepFun66.92026-05-30

08Gemini 3 ProGoogle66.82026-05-17

09GLM-4.7Z.ai66.62026-08-23

10DeepSeek-V3.2DeepSeek652026-09-24

11GPT-5OpenAI652026-06-06

12Claude Opus 4.5Anthropic62.42026-05-17

13Kimi K2 ThinkingMoonshot62.32026-05-30

14Kimi K2.5Moonshot62.32026-05-30

15Qwen3.5 27BAlibaba62.12026-08-23

browsecomp_with_context_manager

VARIANT16 models · best score per model

#ModelVendorScoreVLast seen

01Gemini 3.1 ProGoogle85.92026-05-18

02Claude Opus 4.6Anthropic842026-05-18

03GPT-5.4OpenAI82.72026-05-18

04GLM-5.1Z.ai79.32026-05-18

05GLM-5Z.ai75.92026-05-18

06Kimi K2.5Moonshot74.92026-05-30

07Step 3.5 FlashStepFun692026-05-30

08Claude Opus 4.5Anthropic67.82026-05-17

09DeepSeek-V3.2DeepSeek67.62026-06-13

10GLM-4.7Z.ai67.52026-05-30

11GPT-5.2OpenAI65.82026-05-17

12MiniMax M2.1MiniMax622026-05-30

13Kimi K2 ThinkingMoonshot60.22026-07-09

14Gemini 3 ProGoogle59.22026-06-13

15MiMo V2 FlashXiaomi58.32026-06-13

BrowseComp (context management)

VARIANT15 models · best score per model

#ModelVendorScoreVLast seen

01GPT-5.6 LunaOpenAI842026-07-30

02Qwen3.5 397B A17BAlibaba78.62026-08-24

03Inkling-SmallThinking Machines77.42026-08-25

04InklingThinking Machines77.12026-07-30

05MiniMax M2.7MiniMax76.32026-08-24

06Kimi K2.5Moonshot74.92026-08-24

07DeepSeek-V4-FlashDeepSeek73.22026-08-24

08GPT-5.2OpenAI702026-08-24

09DeepSeek-V3.2DeepSeek67.62026-06-05

10Gemini 3 ProGoogle67.62026-08-24

11Nemotron 3 Ultra 550B A55BNVIDIA632026-08-24

12MiniMax M2.1MiniMax622026-06-05

13Claude Opus 4.5Anthropic59.22026-08-24

14MiniMax M2MiniMax56.92026-06-05

15Claude Sonnet 4.5Anthropic26.12026-06-05

BrowseComp (w/ Ctx)

VARIANT7 models · best score per model

#ModelVendorScoreVLast seen

01GPT-5.6 SolOpenAI89.42026-07-16

02Claude Fable 5Anthropic882026-07-16

03Gemini 3.1 ProGoogle85.92026-07-16

04DeepSeek-V4-ProDeepSeek83.42026-07-16

05Kimi K2.6Moonshot83.22026-07-16

06InklingThinking Machines77.12026-07-16

07Kimi K2.5Moonshot74.92026-07-16

BrowseComp (single-agent)

VARIANT6 models · best score per model

#ModelVendorScoreVLast seen

01Claude Mythos 5Anthropic882026-06-10

02Claude Mythos PreviewAnthropic87.92026-06-12

03GPT-5.5OpenAI84.42026-08-12

04Claude Opus 4.8Anthropic84.32026-06-12

05Claude Opus 4.6Anthropic83.72026-06-03

06Claude Sonnet 4.6Anthropic76.22026-08-12

BrowseComp (Agent Swarm)

VARIANT5 models · best score per model

#ModelVendorScoreVLast seen

01Kimi K2.6Moonshot86.32026-05-03

02Gemini 3.1 ProGoogle85.92026-06-06

03Claude Opus 4.6Anthropic83.72026-06-06

04GPT-5.4OpenAI82.72026-05-03

05Kimi K2.5Moonshot78.42026-06-15

BrowseComp+ (OpenClaw)

VARIANT5 models · best score per model

#ModelVendorScoreVLast seen

01Qwen3.5 9BAlibaba27.22026-08-10

02LFM2.5-2.6BLiquid AI26.92026-08-10

03Qwen3.5 4BAlibaba24.52026-08-10

04Gemma 4 E4BGoogle15.92026-08-10

05Gemma 4 E2BGoogle8.32026-08-10

BrowseComp w/ tools

VARIANT5 models · best score per model

#ModelVendorScoreVLast seen

01Kimi K2 ThinkingMoonshot60.22026-06-12

02GPT-5OpenAI54.92026-06-12

03DeepSeek-V3.2DeepSeek40.12026-06-12

04Claude Sonnet 4.5Anthropic24.12026-06-12

05Kimi K2 InstructMoonshot7.42026-06-12

BrowseComp-ZH w/ tools

VARIANT5 models · best score per model

#ModelVendorScoreVLast seen

01GPT-5OpenAI632026-06-12

02Kimi K2 ThinkingMoonshot62.32026-06-12

03DeepSeek-V3.2DeepSeek47.92026-06-12

04Claude Sonnet 4.5Anthropic42.42026-06-12

05Kimi K2 InstructMoonshot22.22026-06-12

Other tracked prints

THIN BOARDSunder 5 published models each — listed, never hidden

PrintLeaderScoreModels

BrowseComp (multi-agent)Claude Mythos 593.34

BrowseComp-Plus (with tool result clearing)Claude Opus 4.567.63

BrowseComp-Plus (with tool result clearing + memory)Claude Opus 4.572.93

BrowseComp-Plus (auto-truncation)GPT-572.91

BrowseComp w/ ctx managementInkling77.11

BrowseComp-ZH (w/ Context Manager)Step 3.5 Flash73.71

Also tracked, no published surface rows yet: BrowseComp (with context management)

THE FAMILY BOARD · EVERY PRINT LABELED · EVERY SCORE SOURCE-ATTRIBUTED ON ITS OWN LIVE BOARDAll leaderboards The frontier board Models by vendor