VECTOR WIREAI INTELLIGENCE
⌘K
NVDA$1,847+3.2%MSFT$512+1.1%GOOGL$199-0.4%META$728+2.7%AMD$184-1.2%TSM$212+0.6%PLTR$98+4.1%AI IDX4,821+1.9%
PKT
SEEDRefresh Models —Deals —Regulatory —Sources —

Terminal-Bench Hard

100 models tracked

Terminal-Bench Hard - the Hard subset. Structurally distinct from Terminal-Bench 2.0/2.1 (systematic score gaps on 500+ shared models).

#ModelVendorBest scoreRunsLast seen
1GPT-5.3 CodexOpenAI77.322026-05-02
2Gemini 3.1 ProGoogle68.532026-07-07
3GPT-5.6 SolOpenAI65.952026-07-10
4Claude Opus 4.6Anthropic65.432026-06-01
5GPT-5.6 TerraOpenAI62.942026-07-10
6Claude Fable 5Anthropic62.912026-06-10
7GPT-5.2OpenAI62.242026-05-02
8GPT-5.5OpenAI60.652026-06-01
9Claude Sonnet 4.6Anthropic59.142026-05-02
10Claude Opus 4.8Anthropic58.312026-06-01
11GPT-5.4OpenAI57.632026-05-02
12Gemini 3 ProGoogle56.962026-06-13
13Claude Opus 4.7Anthropic54.522026-06-01
14GPT-5.4 miniOpenAI52.332026-06-01
15GLM-5.2 Full Open SourceZ.ai50.812026-06-17
16Qwen3.7 MaxAlibaba50.812026-05-20
17Qwen3.7 Plus PreviewAlibaba4722026-07-07
18Claude Opus 4.5Anthropic4722026-05-06
19Gemini 3.5 FlashGoogle46.242026-07-07
20DeepSeek-V4-ProDeepSeek46.232026-06-29
21Muse SparkMeta45.522026-07-07
22GPT-5.1OpenAI45.532026-06-20
23Kimi K2.7 CodeMoonshot44.722026-07-07
24Qwen3.6 PlusAlibaba43.922026-07-07
25Kimi K2.6Moonshot43.932026-07-07
26Qwen3.6 Max PreviewAlibaba43.912026-05-02
27GLM-5Z.ai43.222026-05-02
28GLM-5.1Z.ai43.222026-05-02
29GPT-5.4 nanoOpenAI42.432026-06-01
30GPT-5.5 InstantOpenAI42.422026-07-07
31MiniMax M3MiniMax42.422026-07-07
32MiMo V2.5 ProXiaomi41.732026-07-07
33Grok 4.20SpaceXAI40.912026-05-02
34Qwen3.5 397B A17BAlibaba40.922026-06-01
35MiMo V2 ProXiaomi40.922026-07-07
36MiniMax M2.7MiniMax39.422026-07-07
37DeepSeek-V4-FlashDeepSeek38.632026-06-01
38Gemini 3 Flash PreviewGoogle38.622026-05-02
39Grok 4SpaceXAI37.922026-07-07
40GPT-5OpenAI37.982026-07-07
41Grok 4.20 0309 V2SpaceXAI37.932026-06-18
42GPT-5 CodexOpenAI37.912026-05-02
43Grok 4.3SpaceXAI37.952026-05-23
44o3OpenAI37.122026-07-07
45GPT-5.2 CodexOpenAI37.112026-05-02
46Gemma 4 31BGoogle36.422026-06-01
47Nemotron 3 Ultra 550B A55BNVIDIA36.412026-06-06
48Claude Sonnet 4.5Anthropic35.642026-06-06
49DeepSeek-V3.2DeepSeek35.662026-06-13
50Step 3.7 FlashStepFun35.622026-07-07
51MiMo V2 OmniXiaomi35.642026-07-07
52Qwen3.6 35B A3BAlibaba34.822026-06-01
53DeepSeek-V3.2-SpecialeDeepSeek34.822026-07-07
54MiniMax M2.5MiniMax34.822026-07-07
55GPT-5.1 CodexOpenAI34.812026-05-02
56Kimi K2.5Moonshot34.822026-06-01
57Qwen3.6 27BAlibaba34.822026-05-02
58Claude Opus 4.1Anthropic34.312026-06-01
59Hy3-previewTencent34.122026-05-02
60GLM-5-TurboZ.ai33.322026-07-07
61GPT-5 miniOpenAI33.352026-07-07
62GPT-5.1 Codex miniOpenAI33.312026-05-02
63GLM-4.7Z.ai33.332026-06-01
64Step 3.5 FlashStepFun32.642026-07-07
65Qwen3.5 27BAlibaba32.622026-05-02
66GLM 5V TurboZ.ai32.612026-06-01
67DeepSeek-V3.1-TerminusDeepSeek31.822026-06-01
68North-Mini-Code-1.0Cohere31.132026-08-23
69Claude Sonnet 4Anthropic31.132026-06-23
70Qwen3.5 122B A10BAlibaba31.122026-05-02
71Kimi K2 (Reasoning)Moonshot31.162026-07-07
72MiMo V2 FlashXiaomi31.162026-07-07
73ling-2-6-1tInclusionAI31.122026-07-07
74DeepSeek-V3.2-ExpDeepSeek31.122026-06-01
75Claude 4 OpusAnthropic31.112026-06-01
76JT-35B-FlashChina Mobile28.822026-07-07
77MiniMax M2.1MiniMax28.822026-07-07
78GLM-4.6Z.ai28.842026-06-19
79Nemotron 3 Super 120B A12BNVIDIA28.812026-06-01
80Claude Haiku 4.5Anthropic27.322026-06-01
81Mercury 2Inception26.522026-07-07
82Gemini 2.5 ProGoogle26.532026-07-07
83Qwen3.5 35B A3BAlibaba26.522026-05-02
84Doubao Seed CodeByteDance26.522026-07-07
85MiniMax M2MiniMax25.832026-07-07
86ERNIE 5.0 Thinking PreviewBaidu2512026-05-02
87DeepSeek-V3.1DeepSeek2522026-06-01
88Gemma 4 26B A4BGoogle2522026-05-06
89Command A+Cohere2532026-08-23
90Qwen3.5 9BAlibaba24.222026-05-02
91Grok 4.1 FastSpaceXAI24.222026-06-23
92Nova 2.0 Pro PreviewAmazon24.232026-05-02
93Gemini 3.1 Flash Lite PreviewGoogle24.222026-07-07
94HyperNova 60B 2605Multiverse Computing23.522026-07-07
95GPT OSS 120BOpenAI23.522026-05-02
96Kimi K2 (Non-Reasoning)Moonshot23.522026-05-02
97K-EXAONELG AI22.722026-05-02
98Trinity Large ThinkingArcee AI22.712026-05-02
99GLM-4.5Z.ai2212026-06-01
100GLM-4.7-FlashZ.ai2222026-06-01

Best tracked score per model (default configuration; source-attributed and verification-tiered). Open a model for its full benchmark surface, provenance and pricing.