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
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 | |
|---|---|---|
| 151 | MiniMax M2.1terminal-bench-2.0, Terminus 2· Terminal-Bench | 29.2% |
| 152 | Claude Haiku 4.5terminal-bench-2.0, Terminus 2· Terminal-Bench | 28.3% |
| 153 | Mistral Large 4terminal-bench-4.0, Claude Code· Terminal-Bench | 27.9% |
| 154 | Kimi K2 Instructterminal-bench-2.0, Terminus 2· Terminal-Bench | 27.8% |
| 155 | Mistral Large 4terminal-bench-4.1, Claude Code· Terminal-Bench | 27.6% |
| 156 | Claude Haiku 4.5terminal-bench-2.0, Claude Code· Terminal-Bench | 27.5% |
| 157 | Grok 4terminal-bench-2.0, OpenHands· Terminal-Bench | 27.2% |
| 158 | Qwen3 Coderterminal-bench-2.0, Dakou Agent· Terminal-Bench | 27.2% |
| 159 | MiMo V2.6 Proterminal-bench-4.0, mini-swe-agent· Terminal-Bench | 27% |
| 160 | MiMo V2.6 Proterminal-bench-4.1, mini-swe-agent· Terminal-Bench | 27% |
| 161 | Qwen3.8 Maxterminal-bench-4.0, Claude Code· Terminal-Bench | 27% |
| 162 | Qwen3.8 Maxterminal-bench-4.1, Claude Code· Terminal-Bench | 27% |
| 163 | Kimi K2 Instructterminal-bench-2.0, OpenHands· Terminal-Bench | 26.7% |
| 164 | Grok 4.6terminal-bench-3.0, Grok Build· Terminal-Bench | 26.5% |
| 165 | Gemini 2.5 Proterminal-bench-2.0, Mini-SWE-Agent· Terminal-Bench | 26.1% |
| 166 | Grok Code Fast 1terminal-bench-2.0, Mini-SWE-Agent· Terminal-Bench | 25.8% |
| 167 | Grok 4terminal-bench-2.0, Mini-SWE-Agent· Terminal-Bench | 25.4% |
| 168 | Qwen3 Coderterminal-bench-2.0, OpenHands· Terminal-Bench | 25.4% |
| 169 | Qwen3.6 35B A3Bterminal-bench-2.0, little-coder· Terminal-Bench | 24.6% |
| 170 | GLM 4.6terminal-bench-2.0, Terminus 2· Terminal-Bench | 24.5% |
| 171 | GPT-5 miniterminal-bench-2.0, Terminus 2· Terminal-Bench | 24% |
| 172 | Qwen3 Coderterminal-bench-2.0, Terminus 2· Terminal-Bench | 23.9% |
| 173 | Claude Opus 4.8terminal-bench-4.0, Claude Code· Terminal-Bench | 23.6% |
| 174 | Claude Opus 4.8terminal-bench-4.1, Claude Code· Terminal-Bench | 23.6% |
| 175 | Grok 4terminal-bench-2.0, Terminus 2· Terminal-Bench | 23.1% |
| 176 | Kimi K3terminal-bench-4.1, Claude Code· Terminal-Bench | 23% |
| 177 | Kimi K3terminal-bench-4.0, Claude Code· Terminal-Bench | 22.4% |
| 178 | GPT-5 miniterminal-bench-2.0, Mini-SWE-Agent· Terminal-Bench | 22.2% |
| 179 | Ember-1terminal-bench-4.0, mini-swe-agent· Terminal-Bench | 21.8% |
| 180 | Ember-1terminal-bench-4.1, mini-swe-agent· Terminal-Bench | 21.8% |
| 181 | GPT-5 nanoterminal-bench-2.0, spoox-o-m· Terminal-Bench | 21.8% |
| 182 | GPT-5.6 Terraterminal-bench-4.0, Codex· Terminal-Bench | 21.5% |
| 183 | GPT-5.6 Terraterminal-bench-4.1, Codex· Terminal-Bench | 21.5% |
| 184 | Claude Opus 4.8terminal-bench-3.0, Claude Code· Terminal-Bench | 21.1% |
| 185 | GPT-5.6 Terraterminal-bench-3.0, Codex· Terminal-Bench | 20.8% |
| 186 | Grok 4.6terminal-bench-4.0, Grok Build· Terminal-Bench | 20.3% |
| 187 | Grok 4.6terminal-bench-4.1, Grok Build· Terminal-Bench | 19.7% |
| 188 | Gemini 2.5 Proterminal-bench-2.0, Gemini CLI· Terminal-Bench | 19.6% |
| 189 | Hy4 previewterminal-bench-4.0, Claude Code· Terminal-Bench | 19.4% |
| 190 | Gemini 3.8 Flashterminal-bench-4.0, mini-SWE-agent· Terminal-Bench | 19.1% |
| 191 | Hy4 previewterminal-bench-4.1, Claude Code· Terminal-Bench | 19.1% |
| 192 | GPT OSS 120Bterminal-bench-2.0, Terminus 2· Terminal-Bench | 18.7% |
| 193 | Gemini 3.8 Flashterminal-bench-4.1, mini-SWE-agent· Terminal-Bench | 18.2% |
| 194 | GPT-5.6 Lunaterminal-bench-4.1, Codex· Terminal-Bench | 17.6% |
| 195 | GPT-5.6 Lunaterminal-bench-4.0, Codex· Terminal-Bench | 17.3% |
| 196 | Gemini 2.5 Flashterminal-bench-2.0, Mini-SWE-Agent· Terminal-Bench | 17.1% |
| 197 | Gemini 2.5 Flashterminal-bench-2.0, Terminus 2· Terminal-Bench | 16.9% |
| 198 | Gemini 2.5 Flashterminal-bench-2.0, OpenHands· Terminal-Bench | 16.4% |
| 199 | Gemini 2.5 Proterminal-bench-2.0, OpenHands· Terminal-Bench | 16.4% |
| 200 | GPT-6 Lunaterminal-bench-4.0, Codex· Terminal-Bench | 16.4% |
Showing 151–200 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)