Skip to content
Weights Not Walls
Menu

Terminal Bench

Reported

Tasks an agent must complete in a command-line terminal, checked by tests.

Kind
Published by a third party's leaderboard, collected by Epoch AI. The leaderboard is named and linked.
Link
https://www.tbench.ai/leaderboard/terminal-bench/2.0
Score shown
Epoch's column "Accuracy mean", as a percentage, higher is better; random guessing scores about 0.0%. Error bars as the source gives them.
Results
202 in all: 35 for open-weight models, 163 for closed ones, 4 where Epoch does not say. 35 linked to a model page here.

Best open-weight result: zai-org/GLM-5, 52.4%, number 67 of 202 results.

Terminal Bench results
#ModelWeightsSettingScoreDateSource
1gpt-5.5_unknownOpenAIclosedeffort unknown; Agent: NexAU-AHE84.7%±2.123 Apr 2026tbench.ai
2gpt-5.5_unknownOpenAIclosedeffort unknown; Agent: Capy83.1%±2.123 Apr 2026tbench.ai
3gpt-5.5_unknownOpenAIclosedeffort unknown; Agent: Codex CLI82.2%±2.223 Apr 2026tbench.ai
4gpt-5.5_unknownOpenAIclosedeffort unknown; Agent: Codex82.0%±2.223 Apr 2026tbench.ai
5gpt-5.4-2026-03-05_unknownOpenAIclosedeffort unknown; Agent: ForgeCode81.8%±2.012 Mar 2026tbench.ai
6claude-opus-4-7_unknownAnthropicclosedeffort unknown; Agent: WOZCODE80.2%±2.116 Apr 2026tbench.ai
7gemini-3.1-pro-previewGoogle DeepMindclosedAgent: TongAgents80.2%±2.619 Feb 2026tbench.ai
8claude-opus-4-6_unknownAnthropicclosedeffort unknown; Agent: ForgeCode79.8%±1.612 Mar 2026tbench.ai
9gpt-5.3-codexOpenAIclosedAgent: SageAgent78.4%±2.25 Feb 2026tbench.ai
10gemini-3.1-pro-previewGoogle DeepMindclosedAgent: Forge Code78.4%±1.82 Mar 2026Terminal-Bench v2 Leaderboard
11gemini-3.1-pro-previewGoogle DeepMindclosedAgent: ForgeCode78.4%±1.82 Mar 2026tbench.ai
12gpt-5.3-codexOpenAIclosedAgent: Droid77.3%±2.25 Feb 2026tbench.ai
13claude-opus-4-6_unknownAnthropicclosedeffort unknown; Agent: Meta-Harness76.4%±2.45 Feb 2026tbench.ai
14gpt-5.3-codexOpenAIclosedAgent: CodeBrain-1.575.8%±2.05 Feb 2026tbench.ai
15gpt-5.3-codexOpenAIclosedAgent: Codelia75.7%±2.25 Feb 2026tbench.ai
16claude-opus-4-6_unknownAnthropicclosedeffort unknown; Agent: Capy75.3%±2.45 Feb 2026tbench.ai
17gpt-5.3-codexOpenAIclosedAgent: Simple Codex75.1%±2.45 Feb 2026tbench.ai
18gemini-3.1-pro-previewGoogle DeepMindclosedAgent: Terminus-KIRA74.8%±2.619 Feb 2026tbench.ai
19claude-opus-4-6_unknownAnthropicclosedeffort unknown; Agent: Terminus-KIRA74.7%±2.65 Feb 2026tbench.ai
20gpt-5.3-codexOpenAIclosedAgent: Mux74.6%±2.55 Feb 2026tbench.ai
21claude-opus-4-6_unknownAnthropicclosedeffort unknown; Agent: MAYA-V272.1%±2.25 Feb 2026tbench.ai
22claude-opus-4-6_unknownAnthropicclosedeffort unknown; Agent: TongAgents71.9%±2.75 Feb 2026tbench.ai
23gpt-5.3-codexOpenAIclosedAgent: spoox-o-m71.5%±2.55 Feb 2026tbench.ai
24gpt-5.3-codexOpenAIclosedAgent: CodeBrain-170.3%±2.610 Feb 2026via Epoch AI
25claude-opus-4-6AnthropicclosedAgent: Droid69.9%±2.55 Feb 2026Terminal-Bench v2 Leaderboard
26claude-opus-4-6_unknownAnthropicclosedeffort unknown; Agent: Droid69.9%±2.55 Feb 2026tbench.ai
27gemini-3-pro-previewGoogle DeepMindclosedAgent: Ante69.4%±2.118 Nov 2025tbench.ai
28gpt-5.3-codexOpenAIclosedAgent: IndusAGI Coding Agent69.1%±2.35 Feb 2026tbench.ai
29claude-opus-4-6_unknownAnthropicclosedeffort unknown; Agent: Crux66.9%5 Feb 2026tbench.ai
30claude-opus-4-6_unknownAnthropicclosedeffort unknown; Agent: Mux66.5%±2.55 Feb 2026tbench.ai
31gpt-5.2-codexOpenAIclosedAgent: Deep Agents66.5%±3.114 Jan 2026tbench.ai
32gpt-5.5_unknownOpenAIclosedeffort unknown; Agent: clnkr66.1%±2.523 Apr 2026tbench.ai
33gemini-3-pro-previewGoogle DeepMindclosedAgent: SageAgent65.2%±2.118 Nov 2025tbench.ai
34gpt-5.2-2025-12-11_unknownOpenAIclosedeffort unknown; Agent: Droid64.9%±2.811 Dec 2025tbench.ai
35gpt-5.2-2025-12-11_mediumOpenAIclosedeffort medium; Agent: Droid64.9%±2.824 Dec 2025Terminal-Bench v2 Leaderboard
36gpt-5.3-codexOpenAIclosedAgent: Terminus 264.7%±2.75 Feb 2026tbench.ai
37gemini-3-flash-previewGoogle DeepMindclosedAgent: Junie CLI64.3%±2.817 Dec 2025tbench.ai
38claude-opus-4-5-20251101_unknownAnthropicclosedeffort unknown; Agent: Droid63.1%±2.724 Nov 2025tbench.ai
39claude-opus-4-5-20251101AnthropicclosedAgent: Droid63.1%±2.711 Dec 2025Terminal-Bench v2 Leaderboard
40claude-opus-4-6_unknownAnthropicclosedeffort unknown; Agent: Terminus 262.9%±2.75 Feb 2026tbench.ai
41gpt-5.2-2025-12-11_unknownOpenAIclosedeffort unknown; Agent: Codex CLI62.9%±3.011 Dec 2025tbench.ai
42gpt-5.2-2025-12-11_mediumOpenAIclosedeffort medium; Agent: Codex CLI62.9%±3.018 Dec 2025Terminal-Bench v2 Leaderboard
43gemini-3-pro-previewGoogle DeepMindclosedAgent: CodeBrain-1.562.2%±2.618 Nov 2025tbench.ai
44gemini-3-pro-previewGoogle DeepMindclosedAgent: CodeBrain-162.2%±2.65 Feb 2026via Epoch AI
45gemini-3-pro-previewGoogle DeepMindclosedAgent: II-Agent61.8%±2.818 Nov 2025tbench.ai
46gpt-5.1-codex-miniOpenAIclosedAgent: hookele61.6%±1.912 Nov 2025tbench.ai
47gemini-3.1-pro-previewGoogle DeepMindclosedAgent: Gemini CLI61.4%±4.119 Feb 2026tbench.ai
48gemini-3-pro-previewGoogle DeepMindclosedAgent: Droid61.1%±2.818 Nov 2025tbench.ai
49gpt-5.2-2025-12-11_unknownOpenAIclosedeffort unknown; Agent: Mux60.7%11 Dec 2025tbench.ai
50gpt-5.1-codex-maxOpenAIclosedAgent: Codex CLI60.4%±2.719 Nov 2025tbench.ai
51claude-opus-4-5-20251101_unknownAnthropicclosedeffort unknown; Agent: Letta Code59.1%±2.424 Nov 2025tbench.ai
52claude-opus-4-5-20251101_128KAnthropicclosedAgent: Letta Code59.1%±2.417 Dec 2025Terminal-Bench v2 Leaderboard
53claude-opus-4-5-20251101_unknownAnthropicclosedeffort unknown; Agent: Mux58.4%24 Nov 2025tbench.ai
54claude-opus-4-6_unknownAnthropicclosedeffort unknown; Agent: Claude Code58.0%±2.95 Feb 2026tbench.ai
55gpt-5.1-codexOpenAIclosedAgent: Codex CLI57.8%±2.916 Nov 2025Terminal-Bench v2 Leaderboard
56claude-opus-4-5-20251101_unknownAnthropicclosedeffort unknown; Agent: Terminus 257.8%±2.524 Nov 2025tbench.ai
57gpt-5.1-codexOpenAIclosedAgent: Crux57.8%±2.912 Nov 2025tbench.ai
58grok-4-20xAIclosedAgent: Grok CLI57.3%10 Mar 2026tbench.ai
59grok-4-20xAIclosedAgent: grok-cli57.3%2 Apr 2026tbench.ai
60gemini-3-pro-previewGoogle DeepMindclosedAgent: Terminus 256.9%±2.518 Nov 2025tbench.ai
61gemini-3-pro-previewGoogle DeepMindclosedAgent: Letta Code56.0%±3.018 Nov 2025tbench.ai
62claude-opus-4-5-20251101_unknownAnthropicclosedeffort unknown; Agent: Goose54.3%±2.624 Nov 2025tbench.ai
63gpt-5.2-2025-12-11_unknownOpenAIclosedeffort unknown; Agent: Terminus 254.0%±2.911 Dec 2025tbench.ai
64gpt-5.1-codexOpenAIclosedAgent: Letta Code53.5%±2.812 Nov 2025tbench.ai
65claude-sonnet-4-6_unknownAnthropicclosedeffort unknown; Agent: Simplai Agent53.4%±2.817 Feb 2026tbench.ai
66claude-sonnet-4-6AnthropicclosedAgent: Simplai Agent53.4%±2.814 May 2026via Epoch AI
67zai-org/GLM-5Epoch: glm-5, Z.ai (Zhipu AI)openAgent: Terminus 252.4%±2.611 Feb 2026tbench.ai
68claude-opus-4-5-20251101_unknownAnthropicclosedeffort unknown; Agent: Claude Code52.1%±2.524 Nov 2025tbench.ai
69claude-opus-4-5-20251101_unknownAnthropicclosedeffort unknown; Agent: OpenHands51.9%±2.924 Nov 2025tbench.ai
70claude-opus-4-5-20251101_unknownAnthropicclosedeffort unknown; Agent: OpenCode51.7%24 Nov 2025tbench.ai
71gemini-3-flash-previewGoogle DeepMindclosedAgent: Terminus 251.7%±3.117 Dec 2025tbench.ai
72gemini-3-flash-previewGoogle DeepMindclosedAgent: Gemini CLI51.0%±3.023 Dec 2025Terminal-Bench v2 Leaderboard
73gpt-5-2025-08-07_unknownOpenAIclosedeffort unknown; Agent: Codex CLI49.6%±2.97 Aug 2025tbench.ai
74gpt-5-2025-08-07_mediumOpenAIclosedeffort medium; Agent: Codex CLI49.6%±2.94 Nov 2025Terminal-Bench v2 Leaderboard
75gpt-5.1-2025-11-13_mediumOpenAIclosedeffort medium; Agent: Terminus 247.6%±2.816 Nov 2025Terminal-Bench v2 Leaderboard
76gpt-5.1-2025-11-13_mediumOpenAIclosedeffort medium; Agent: Terminus 247.6%±2.816 Nov 2025Terminal-Bench v2 Leaderboard
77gpt-5.1-2025-11-13_unknownOpenAIclosedeffort unknown; Agent: Terminus 247.6%±2.812 Nov 2025tbench.ai
78gemini-3-flash-previewGoogle DeepMindclosedAgent: Gemini CLI47.4%±3.017 Dec 2025tbench.ai
79claude-sonnet-4-5-20250929_unknownAnthropicclosedeffort unknown; Agent: CAMEL-AI46.5%±2.429 Sep 2025tbench.ai
80MiniMaxAI/MiniMax-M2.7Epoch: MiniMax-M2.7, MiniMaxopenAgent: IndusAGI Coding Agent45.1%18 Mar 2026tbench.ai
81gpt-5-codexOpenAIclosedAgent: Codex CLI44.3%±2.715 Sep 2025tbench.ai
82gpt-5-codexOpenAIclosedAgent: Codex CLI44.3%±2.74 Nov 2025Terminal-Bench v2 Leaderboard
83gpt-5-2025-08-07_unknownOpenAIclosedeffort unknown; Agent: OpenHands43.8%±3.07 Aug 2025tbench.ai
84gpt-5-2025-08-07_mediumOpenAIclosedeffort medium; Agent: OpenHands43.8%±3.02 Nov 2025Terminal-Bench v2 Leaderboard
85gpt-5-codexOpenAIclosedAgent: Terminus 243.4%±2.915 Sep 2025tbench.ai
86gpt-5-codexOpenAIclosedAgent: Terminus 243.4%±2.931 Oct 2025Terminal-Bench v2 Leaderboard
87moonshotai/Kimi-K2.5Epoch: kimi-k2.5, MoonshotopenAgent: Terminus 243.2%±2.927 Jan 2026tbench.ai
88claude-sonnet-4-5-20250929_unknownAnthropicclosedeffort unknown; Agent: Goose43.1%±2.629 Sep 2025tbench.ai
89gpt-5.1-codex-miniOpenAIclosedAgent: Crux43.1%±3.012 Nov 2025tbench.ai
90gpt-5.1-codex-miniOpenAIclosedAgent: Codex CLI43.1%±3.017 Nov 2025Terminal-Bench v2 Leaderboard
91MiniMaxAI/MiniMax-M2.7Epoch: MiniMax-M2.7, MiniMaxopenAgent: Harness Agent42.9%±2.918 Mar 2026tbench.ai
92claude-sonnet-4-5-20250929_unknownAnthropicclosedeffort unknown; Agent: Terminus 242.8%±2.829 Sep 2025tbench.ai
93claude-sonnet-4-5-20250929AnthropicclosedAgent: Terminus 242.8%±2.831 Oct 2025Terminal-Bench v2 Leaderboard
94claude-sonnet-4-5-20250929_unknownAnthropicclosedeffort unknown; Agent: MAYA42.7%4 Jan 2026via Epoch AI
95claude-sonnet-4-5-20250929_unknownAnthropicclosedeffort unknown; Agent: MAYA-V242.7%29 Sep 2025tbench.ai
96MiniMaxAI/MiniMax-M2.5Epoch: MiniMax-M2.5, MiniMaxopenAgent: cchuter42.7%±2.812 Feb 2026tbench.ai
97claude-sonnet-4-5-20250929AnthropicclosedAgent: OpenHands42.6%±2.82 Nov 2025Terminal-Bench v2 Leaderboard
98claude-sonnet-4-5-20250929_unknownAnthropicclosedeffort unknown; Agent: OpenHands42.6%±2.829 Sep 2025tbench.ai
99claude-sonnet-4-5-20250929_unknownAnthropicclosedeffort unknown; Agent: Mini-SWE-Agent42.5%±2.829 Sep 2025tbench.ai
100claude-sonnet-4-5-20250929AnthropicclosedAgent: Mini-SWE-Agent42.5%±2.83 Nov 2025Terminal-Bench v2 Leaderboard
101MiniMaxAI/MiniMax-M2.5Epoch: MiniMax-M2.5, MiniMaxopenAgent: Terminus 242.2%±2.612 Feb 2026tbench.ai
102gpt-5-codexOpenAIclosedAgent: Mini-SWE-Agent41.3%±2.815 Sep 2025tbench.ai
103gpt-5-codexOpenAIclosedAgent: Mini-SWE-Agent41.3%±2.83 Nov 2025Terminal-Bench v2 Leaderboard
104claude-sonnet-4-5-20250929AnthropicclosedAgent: Claude Code40.1%±2.94 Nov 2025Terminal-Bench v2 Leaderboard
105claude-sonnet-4-5-20250929_unknownAnthropicclosedeffort unknown; Agent: Claude Code40.1%±2.929 Sep 2025tbench.ai
106deepseek-ai/DeepSeek-V3.2Epoch: deepseek/deepseek-v3.2, DeepSeekopenAgent: Terminus 239.6%±2.81 Dec 2025tbench.ai
107claude-opus-4-1-20250805AnthropicclosedAgent: Terminus 238.0%±2.631 Oct 2025Terminal-Bench v2 Leaderboard
108claude-opus-4-1-20250805_unknownAnthropicclosedeffort unknown; Agent: Terminus 238.0%±2.65 Aug 2025tbench.ai
109claude-opus-4-1-20250805AnthropicclosedAgent: OpenHands36.9%±2.72 Nov 2025Terminal-Bench v2 Leaderboard
110gpt-5.1-codexOpenAIclosedAgent: Terminus 236.9%±3.217 Nov 2025Terminal-Bench v2 Leaderboard
111claude-opus-4-1-20250805_unknownAnthropicclosedeffort unknown; Agent: OpenHands36.9%±2.75 Aug 2025tbench.ai
112gpt-5.1-codexOpenAIclosedAgent: Terminus 236.9%±3.212 Nov 2025tbench.ai
113MiniMaxAI/MiniMax-M2.1Epoch: MiniMax-M2.1, MiniMaxopenAgent: Crux36.6%±2.923 Dec 2025tbench.ai
114moonshotai/Kimi-K2-ThinkingEpoch: kimi-k2-thinking, MoonshotopenAgent: Terminus 235.7%±2.86 Nov 2025tbench.ai
115moonshotai/Kimi-K2-ThinkingEpoch: kimi-k2-thinking, MoonshotopenAgent: Terminus 235.7%±2.811 Nov 2025Terminal-Bench v2 Leaderboard
116claude-haiku-4-5-20251001_unknownAnthropicclosedeffort unknown; Agent: Goose35.5%±2.915 Oct 2025tbench.ai
117gpt-5-2025-08-07_mediumOpenAIclosedeffort medium; Agent: Terminus 235.2%±3.131 Oct 2025Terminal-Bench v2 Leaderboard
118gpt-5-2025-08-07_unknownOpenAIclosedeffort unknown; Agent: Terminus 235.2%±3.17 Aug 2025tbench.ai
119claude-opus-4-1-20250805AnthropicclosedAgent: Mini-SWE-Agent35.1%±2.53 Nov 2025Terminal-Bench v2 Leaderboard
120claude-opus-4-1-20250805_unknownAnthropicclosedeffort unknown; Agent: Mini-SWE-Agent35.1%±2.55 Aug 2025tbench.ai
121claude-opus-4-1-20250805_unknownAnthropicclosedeffort unknown; Agent: Claude Code34.8%±2.95 Aug 2025tbench.ai
122gpt-5-mini-2025-08-07_unknownOpenAIclosedeffort unknown; Agent: spoox-o-m34.8%±2.77 Aug 2025tbench.ai
123claude-opus-4-1-20250805AnthropicclosedAgent: Claude Code34.8%±2.94 Nov 2025Terminal-Bench v2 Leaderboard
124gpt-5-mini-2025-08-07_unknownOpenAIclosedeffort unknown; Agent: spoox-m34.8%±2.724 Dec 2025via Epoch AI
125gpt-5-2025-08-07_unknownOpenAIclosedeffort unknown; Agent: Mini-SWE-Agent33.9%±2.97 Aug 2025tbench.ai
126gpt-5-2025-08-07_mediumOpenAIclosedeffort medium; Agent: Mini-SWE-Agent33.9%±2.93 Nov 2025Terminal-Bench v2 Leaderboard
127zai-org/GLM-4.7Epoch: glm-4.7, Z.ai (Zhipu AI)openAgent: Terminus 233.4%±2.822 Dec 2025tbench.ai
128zai-org/GLM-4.7Epoch: glm-4.7, Z.ai (Zhipu AI)openAgent: Crux33.3%±2.522 Dec 2025tbench.ai
129gemini-2.5-proGoogle DeepMindclosedAgent: Terminus 232.6%±3.017 Jun 2025tbench.ai
130gemini-2.5-proGoogle DeepMindclosedAgent: Terminus 232.6%±3.031 Oct 2025Terminal-Bench v2 Leaderboard
131gpt-5-mini-2025-08-07_mediumOpenAIclosedeffort medium; Agent: Codex CLI31.9%±3.04 Nov 2025Terminal-Bench v2 Leaderboard
132gpt-5-mini-2025-08-07_unknownOpenAIclosedeffort unknown; Agent: Codex CLI31.9%±3.07 Aug 2025tbench.ai
133MiniMaxAI/MiniMax-M2Epoch: MiniMax-M2, MiniMaxopenAgent: Terminus 230.0%±2.727 Oct 2025tbench.ai
134claude-haiku-4-5-20251001_unknownAnthropicclosedeffort unknown; Agent: Mini-SWE-Agent29.8%±2.515 Oct 2025tbench.ai
135claude-haiku-4-5-20251001AnthropicclosedAgent: Mini-SWE-Agent29.8%±2.53 Nov 2025Terminal-Bench v2 Leaderboard
136gpt-5-mini-2025-08-07_unknownOpenAIclosedeffort unknown; Agent: OpenHands29.2%±2.87 Aug 2025tbench.ai
137MiniMaxAI/MiniMax-M2.1Epoch: MiniMax-M2.1, MiniMaxopenAgent: Terminus 229.2%±2.923 Dec 2025tbench.ai
138gpt-5-mini-2025-08-07_mediumOpenAIclosedeffort medium; Agent: OpenHands29.2%±2.82 Nov 2025Terminal-Bench v2 Leaderboard
139claude-haiku-4-5-20251001_unknownAnthropicclosedeffort unknown; Agent: Terminus 228.3%±2.915 Oct 2025tbench.ai
140claude-haiku-4-5-20251001AnthropicclosedAgent: Terminus 228.3%±2.931 Oct 2025Terminal-Bench v2 Leaderboard
141moonshotai/Kimi-K2-InstructEpoch: Kimi-K2-Instruct, MoonshotopenAgent: Terminus 227.8%±2.51 Nov 2025Terminal-Bench v2 Leaderboard
142moonshotai/Kimi-K2-InstructEpoch: Kimi-K2-Instruct, MoonshotopenAgent: Terminus 227.8%±2.511 Jul 2025tbench.ai
143claude-haiku-4-5-20251001_unknownAnthropicclosedeffort unknown; Agent: Claude Code27.5%±2.815 Oct 2025tbench.ai
144claude-haiku-4-5-20251001AnthropicclosedAgent: Claude Code27.5%±2.84 Nov 2025Terminal-Bench v2 Leaderboard
145grok-4-0709xAIclosedAgent: OpenHands27.2%±3.12 Nov 2025Terminal-Bench v2 Leaderboard
146grok-4-0709xAIclosedAgent: OpenHands27.2%±3.19 Jul 2025tbench.ai
147Qwen/Qwen3-Coder-480B-A35B-InstructEpoch: Qwen3-Coder-480B-A35B-Instruct, AlibabaopenAgent: Dakou Agent27.2%±2.622 Jul 2025tbench.ai
148moonshotai/Kimi-K2-InstructEpoch: Kimi-K2-Instruct, MoonshotopenAgent: OpenHands26.7%±2.711 Jul 2025tbench.ai
149moonshotai/Kimi-K2-InstructEpoch: Kimi-K2-Instruct, MoonshotopenAgent: OpenHands26.7%±2.72 Nov 2025Terminal-Bench v2 Leaderboard
150gemini-2.5-proGoogle DeepMindclosedAgent: Mini-SWE-Agent26.1%±2.53 Nov 2025Terminal-Bench v2 Leaderboard
151gemini-2.5-proGoogle DeepMindclosedAgent: Mini-SWE-Agent26.1%±2.517 Jun 2025tbench.ai
152grok-code-fast-1not statedAgent: Mini-SWE-Agent25.8%±2.628 Aug 2025tbench.ai
153grok-code-fast-1not statedAgent: Mini-SWE-Agent25.8%±2.63 Nov 2025Terminal-Bench v2 Leaderboard
154grok-4-0709xAIclosedAgent: Mini-SWE-Agent25.4%±2.93 Nov 2025Terminal-Bench v2 Leaderboard
155Qwen/Qwen3-Coder-480B-A35B-InstructEpoch: Qwen3-Coder-480B-A35B-Instruct, AlibabaopenAgent: OpenHands25.4%±2.62 Nov 2025Terminal-Bench v2 Leaderboard
156grok-4-0709xAIclosedAgent: Mini-SWE-Agent25.4%±2.99 Jul 2025tbench.ai
157Qwen/Qwen3-Coder-480B-A35B-InstructEpoch: Qwen3-Coder-480B-A35B-Instruct, AlibabaopenAgent: OpenHands25.4%±2.622 Jul 2025tbench.ai
158Qwen/Qwen3.6-35B-A3BEpoch: qwen3.6-35b-a3b, AlibabaopenAgent: little-coder24.6%±3.216 Apr 2026tbench.ai
159zai-org/GLM-4.6Epoch: glm-4.6, Z.ai (Zhipu AI),Tsinghua UniversityopenAgent: Terminus 224.5%±2.430 Sep 2025tbench.ai
160zai-org/GLM-4.6Epoch: glm-4.6, Z.ai (Zhipu AI),Tsinghua UniversityopenAgent: Terminus 224.5%±2.41 Nov 2025Terminal-Bench v2 Leaderboard
161gpt-5-mini-2025-08-07_unknownOpenAIclosedeffort unknown; Agent: Terminus 224.0%±2.57 Aug 2025tbench.ai
162gpt-5-mini-2025-08-07_mediumOpenAIclosedeffort medium; Agent: Terminus 224.0%±2.531 Oct 2025Terminal-Bench v2 Leaderboard
163Qwen/Qwen3-Coder-480B-A35B-InstructEpoch: Qwen3-Coder-480B-A35B-Instruct, AlibabaopenAgent: Terminus 223.9%±2.81 Nov 2025Terminal-Bench v2 Leaderboard
164Qwen/Qwen3-Coder-480B-A35B-InstructEpoch: Qwen3-Coder-480B-A35B-Instruct, AlibabaopenAgent: Terminus 223.9%±2.822 Jul 2025tbench.ai
165grok-4-0709xAIclosedAgent: Terminus 223.1%±2.99 Jul 2025tbench.ai
166grok-4-0709xAIclosedAgent: Terminus 223.1%±2.931 Oct 2025Terminal-Bench v2 Leaderboard
167gpt-5-mini-2025-08-07_unknownOpenAIclosedeffort unknown; Agent: Mini-SWE-Agent22.2%±2.67 Aug 2025tbench.ai
168gpt-5-mini-2025-08-07_mediumOpenAIclosedeffort medium; Agent: Mini-SWE-Agent22.2%±2.63 Nov 2025Terminal-Bench v2 Leaderboard
169gpt-5-nano-2025-08-07_unknownOpenAIclosedeffort unknown; Agent: spoox-o-m21.8%±2.87 Aug 2025tbench.ai
170gemini-2.5-proGoogle DeepMindclosedAgent: Gemini CLI19.6%±2.94 Nov 2025Terminal-Bench v2 Leaderboard
171gemini-2.5-proGoogle DeepMindclosedAgent: Gemini CLI19.6%±2.917 Jun 2025tbench.ai
172openai/gpt-oss-120bEpoch: gpt-oss-120b, OpenAIopenAgent: Terminus 218.7%±2.71 Nov 2025Terminal-Bench v2 Leaderboard
173openai/gpt-oss-120bEpoch: gpt-oss-120b_unknown, OpenAIopeneffort unknown; Agent: Terminus 218.7%±2.75 Aug 2025tbench.ai
174gemini-2.5-flash-preview-09-2025Google DeepMindclosedAgent: Mini-SWE-Agent17.1%±2.53 Nov 2025Terminal-Bench v2 Leaderboard
175gemini-2.5-flashGoogle DeepMindclosedAgent: Mini-SWE-Agent17.1%±2.517 Jun 2025tbench.ai
176gemini-2.5-flash-preview-09-2025Google DeepMindclosedAgent: Terminus 216.9%±2.431 Oct 2025Terminal-Bench v2 Leaderboard
177gemini-2.5-flashGoogle DeepMindclosedAgent: Terminus 216.9%±2.417 Jun 2025tbench.ai
178gemini-2.5-flashGoogle DeepMindclosedAgent: OpenHands16.4%±2.417 Jun 2025tbench.ai
179gemini-2.5-proGoogle DeepMindclosedAgent: OpenHands16.4%±2.817 Jun 2025tbench.ai
180gemini-2.5-flash-preview-09-2025Google DeepMindclosedAgent: OpenHands16.4%±2.42 Nov 2025Terminal-Bench v2 Leaderboard
181gemini-2.5-proGoogle DeepMindclosedAgent: OpenHands16.4%±2.82 Nov 2025Terminal-Bench v2 Leaderboard
182gemini-2.5-flashGoogle DeepMindclosedAgent: Gemini CLI15.4%±2.317 Jun 2025tbench.ai
183gemini-2.5-flash-preview-09-2025Google DeepMindclosedAgent: Gemini CLI15.4%±2.34 Nov 2025Terminal-Bench v2 Leaderboard
184openai/gpt-oss-120bEpoch: gpt-oss-120b, OpenAIopenAgent: Mini-SWE-Agent14.2%±2.33 Nov 2025Terminal-Bench v2 Leaderboard
185grok-code-fast-1not statedAgent: Terminus 214.2%±2.531 Oct 2025Terminal-Bench v2 Leaderboard
186openai/gpt-oss-120bEpoch: gpt-oss-120b_unknown, OpenAIopeneffort unknown; Agent: Mini-SWE-Agent14.2%±2.35 Aug 2025tbench.ai
187grok-code-fast-1not statedAgent: Terminus 214.2%±2.528 Aug 2025tbench.ai
188claude-haiku-4-5-20251001_unknownAnthropicclosedeffort unknown; Agent: OpenHands13.9%±2.715 Oct 2025tbench.ai
189claude-haiku-4-5-20251001AnthropicclosedAgent: OpenHands13.9%±2.72 Nov 2025Terminal-Bench v2 Leaderboard
190gpt-5-nano-2025-08-07_mediumOpenAIclosedeffort medium; Agent: Codex CLI11.5%±2.34 Nov 2025Terminal-Bench v2 Leaderboard
191gpt-5-nano-2025-08-07_unknownOpenAIclosedeffort unknown; Agent: Codex CLI11.5%±2.37 Aug 2025tbench.ai
192gpt-5-nano-2025-08-07_mediumOpenAIclosedeffort medium; Agent: OpenHands9.9%±2.12 Nov 2025Terminal-Bench v2 Leaderboard
193gpt-5-nano-2025-08-07_unknownOpenAIclosedeffort unknown; Agent: OpenHands9.9%±2.17 Aug 2025tbench.ai
194Qwen/Qwen3.5-9BEpoch: qwen3.5-9B, AlibabaopenAgent: little-coder9.2%±2.42 Mar 2026tbench.ai
195gpt-5-nano-2025-08-07_mediumOpenAIclosedeffort medium; Agent: Terminus 27.9%±1.931 Oct 2025Terminal-Bench v2 Leaderboard
196gpt-5-nano-2025-08-07_unknownOpenAIclosedeffort unknown; Agent: Terminus 27.9%±1.97 Aug 2025tbench.ai
197gpt-5-nano-2025-08-07_mediumOpenAIclosedeffort medium; Agent: Mini-SWE-Agent7.0%±1.93 Nov 2025Terminal-Bench v2 Leaderboard
198gpt-5-nano-2025-08-07_unknownOpenAIclosedeffort unknown; Agent: Mini-SWE-Agent7.0%±1.97 Aug 2025tbench.ai
199openai/gpt-oss-20bEpoch: gpt-oss-20b, OpenAIopenAgent: Mini-SWE-Agent3.4%±1.43 Nov 2025Terminal-Bench v2 Leaderboard
200openai/gpt-oss-20bEpoch: gpt-oss-20b_unknown, OpenAIopeneffort unknown; Agent: Mini-SWE-Agent3.4%±1.45 Aug 2025tbench.ai
201openai/gpt-oss-20bEpoch: gpt-oss-20b, OpenAIopenAgent: Terminus 23.1%±1.51 Nov 2025Terminal-Bench v2 Leaderboard
202openai/gpt-oss-20bEpoch: gpt-oss-20b_unknown, OpenAIopeneffort unknown; Agent: Terminus 23.1%±1.55 Aug 2025tbench.ai

Open-weight rows are shaded. Dates are when the result was run or evaluated, where the data says. Data from Epoch AI (CC-BY 4.0), fetched 19 Nov 2026: Epoch AI, ‘Capabilities & benchmarking’. Published online at epoch.ai. Retrieved from ‘https://epoch.ai/benchmarks’ [online resource]. Epoch AI. Method