Skip to content

Benchmark

TBench

Agentic benchmarks measure how well models carry out tasks that take several steps. Scores are percentages.

Models measured
89
Scores recorded
234
Sources
1
IQ weight
0.93

Score distribution

42.9%
Mean
42.2%
Median
22.9%
Std dev
14.2%
10th percentile
75.1%
90th percentile
0%
Min
0%91%
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 0.93 for chat models, measured over 89 models. The weight is how much it counts toward llminfo IQ. We measure it from the data, not from the raw scores.

Leaderboard

#Model
51GPT 5.2terminal-bench-2.0, Codex CLI· Terminal-Bench62.9%
52Gemini 3 Pro Previewterminal-bench-2.0, CodeBrain-1.5· Terminal-Bench62.2%
53GPT-6.2 Astraterminal-bench-4.1, Codex· Terminal-Bench62.1%
54Gemini 3 Pro Previewterminal-bench-2.0, II-Agent· Terminal-Bench61.8%
55GPT-5.1 Codex Miniterminal-bench-2.0, hookele· Terminal-Bench61.6%
56GPT-6.2 Solterminal-bench-4.1, Codex· Terminal-Bench61.5%
57Gemini 3.1 Pro Previewterminal-bench-2.0, Gemini CLI· Terminal-Bench61.4%
58Gemini 4 Neonterminal-bench-4.1, Antigravity CLI· Terminal-Bench61.2%
59Gemini 3 Pro Previewterminal-bench-2.0, Droid· Terminal-Bench61.1%
60GPT-6.1 Solterminal-bench-4.1, Codex· Terminal-Bench60.9%
61Kimi K3.1 Proterminal-bench-4.0, Claude Code· Terminal-Bench60.9%
62GPT 5.2terminal-bench-2.0, Mux· Terminal-Bench60.7%
63GPT-5.1 Codex Maxterminal-bench-2.0, Codex CLI· Terminal-Bench60.4%
64GPT-6 Astraterminal-bench-4.1, Codex· Terminal-Bench60.3%
65Kimi K3.1 Proterminal-bench-4.1, Claude Code· Terminal-Bench59.7%
66Claude Opus 4.5terminal-bench-2.0, Letta Code· Terminal-Bench59.1%
67GLM 5.1terminal-bench-2.1, Claude Code· Terminal-Bench58.7%
68Claude Opus 4.5terminal-bench-2.0, Mux· Terminal-Bench58.4%
69GPT-6 Astraterminal-bench-4.0, Codex· Terminal-Bench58.2%
70GPT-6.1 Solterminal-bench-4.0, Codex· Terminal-Bench58.2%
71Claude Opus 4.6terminal-bench-2.0, Claude Code· Terminal-Bench58%
72Claude Opus 4.5terminal-bench-2.0, Terminus 2· Terminal-Bench57.8%
73GPT-5.1 Codexterminal-bench-2.0, Crux· Terminal-Bench57.8%
74Grok 4.20terminal-bench-2.0, Grok CLI· Terminal-Bench57.3%
75Gemini 3 Pro Previewterminal-bench-2.0, Terminus 2· Terminal-Bench56.9%
76Gemini 4 Hydrogenterminal-bench-4.1, Antigravity CLI· Terminal-Bench56.4%
77Gemini 3 Pro Previewterminal-bench-2.0, Letta Code· Terminal-Bench56%
78Gemini 4 Argonterminal-bench-4.1, Antigravity CLI· Terminal-Bench55.5%
79Gemini 4 Argonterminal-bench-4.0, Antigravity CLI· Terminal-Bench54.9%
80Claude Opus 4.5terminal-bench-2.0, Goose· Terminal-Bench54.3%
81GPT 5.2terminal-bench-2.0, Terminus 2· Terminal-Bench54%
82GPT-5.1 Codexterminal-bench-2.0, Letta Code· Terminal-Bench53.5%
83Claude Sonnet 4.6terminal-bench-2.0, Simplai Agent· Terminal-Bench53.4%
84GPT-6.1 Terraterminal-bench-4.1, Codex· Terminal-Bench53.3%
85GLM 5terminal-bench-2.0, Terminus 2· Terminal-Bench52.4%
86Claude Opus 4.5terminal-bench-2.0, Claude Code· Terminal-Bench52.1%
87Claude Opus 4.5terminal-bench-2.0, OpenHands· Terminal-Bench51.9%
88GPT-6.1 Terraterminal-bench-4.0, Codex· Terminal-Bench51.8%
89Claude Opus 4.5terminal-bench-2.0, OpenCode· Terminal-Bench51.7%
90Gemini 3 Flashterminal-bench-2.0, Terminus 2· Terminal-Bench51.7%
91GPT-6 Solterminal-bench-4.1, Codex· Terminal-Bench50.3%
92GPT 5terminal-bench-2.0, Codex CLI· Terminal-Bench49.6%
93GPT-6 Solterminal-bench-4.0, Codex· Terminal-Bench49.4%
94GPT 5.1terminal-bench-2.0, Terminus 2· Terminal-Bench47.6%
95Gemini 3 Flashterminal-bench-2.0, Gemini CLI· Terminal-Bench47.4%
96Muse Wonderterminal-bench-4.0, Muse Code· Terminal-Bench47%
97Muse Wonderterminal-bench-4.1, Muse Code· Terminal-Bench47%
98Claude Sonnet 4.5terminal-bench-2.0, CAMEL-AI· Terminal-Bench46.5%
99MiniMax M2.7terminal-bench-2.0, IndusAGI Coding Agent· Terminal-Bench45.1%
100GPT-5 Codexterminal-bench-2.0, Codex CLI· Terminal-Bench44.3%

Showing 51–100 of 234 results.

Tokens the model generated per task, as measured by the source. Comparable within this benchmark only.

Data credits: Terminal-Bench (Laude Institute), Apache-2.0 (licence)