Terminal Bench
Terminal-Bench results from its public leaderboard that Epoch AI republishes. Each row is a model run in an agent harness.
No filters. Showing 202 of 202 rows.
1[1–6] OpenAIDossier | No effort set | NexAU-AHE | GPT-5.5, Accuracy: 84.7% (scale 0.0 to 100.0), 95% confidence interval 80.6 to 88.884.7%(80.6–88.8) | Runs unstated | Cost kind unknown |
2[1–10] OpenAIDossierTied #1 | No effort set | Capy | GPT-5.5, Accuracy: 83.1% (scale 0.0 to 100.0), 95% confidence interval 79.0 to 87.383.1%(79.0–87.3) | Runs unstated | Cost kind unknown |
3[1–11] OpenAIDossierTied #1 | No effort set | Codex CLI | GPT-5.5, Accuracy: 82.2% (scale 0.0 to 100.0), 95% confidence interval 77.9 to 86.682.2%(77.9–86.6) | Runs unstated | Cost kind unknown |
4[1–12] OpenAIDossierTied #1 | No effort set | Codex CLI | GPT-5.5, Accuracy: 82% (scale 0.0 to 100.0), 95% confidence interval 77.7 to 86.382%(77.7–86.3) | Runs unstated | Cost kind unknown |
5[1–11] OpenAIDossierTied #1 | No effort set | ForgeCode | GPT-5.4, Accuracy: 81.8% (scale 0.0 to 100.0), 95% confidence interval 77.9 to 85.781.8%(77.9–85.7) | Runs unstated | Cost kind unknown |
6[2–16] GoogleDossierTied #1 | No effort set | TongAgents | Gemini 3.1 Pro Preview, Accuracy: 80.2% (scale 0.0 to 100.0), 95% confidence interval 75.1 to 85.480.2%(75.1–85.4) | Runs unstated | Cost kind unknown |
6[2–15] AnthropicDossierTied #1 | No effort set | WOZCODE | Opus 4.7, Accuracy: 80.2% (scale 0.0 to 100.0), 95% confidence interval 76.1 to 84.480.2%(76.1–84.4) | Runs unstated | Cost kind unknown |
8[3–14] AnthropicDossierTied #1 | No effort set | ForgeCode | Opus 4.6, Accuracy: 79.8% (scale 0.0 to 100.0), 95% confidence interval 76.7 to 82.979.8%(76.7–82.9) | Runs unstated | Cost kind unknown |
9[3–19] OpenAIDossier | No effort set | SageAgent | GPT-5.3 Codex, Accuracy: 78.4% (scale 0.0 to 100.0), 95% confidence interval 74.1 to 82.778.4%(74.1–82.7) | Runs unstated | Cost kind unknown |
10[5–17] GoogleDossier | No effort set | Forge Code | Gemini 3.1 Pro Preview, Accuracy: 78.4% (scale 0.0 to 100.0), 95% confidence interval 74.9 to 81.978.4%(74.9–81.9) | Runs unstated | Cost kind unknown |
10[4–17] GoogleDossier | No effort set | ForgeCode | Gemini 3.1 Pro Preview, Accuracy: 78.4% (scale 0.0 to 100.0), 95% confidence interval 74.9 to 81.978.4%(74.9–81.9) | Runs unstated | Cost kind unknown |
12[5–20] OpenAIDossier | No effort set | Droid | GPT-5.3 Codex, Accuracy: 77.3% (scale 0.0 to 100.0), 95% confidence interval 73.1 to 81.577.3%(73.1–81.5) | Runs unstated | Cost kind unknown |
13[6–22] AnthropicDossier | No effort set | Meta-Harness | Opus 4.6, Accuracy: 76.4% (scale 0.0 to 100.0), 95% confidence interval 71.6 to 81.276.4%(71.6–81.2) | Runs unstated | Cost kind unknown |
14[8–23] OpenAIDossier | No effort set | CodeBrain-1.5 | GPT-5.3 Codex, Accuracy: 75.8% (scale 0.0 to 100.0), 95% confidence interval 71.9 to 79.875.8%(71.9–79.8) | Runs unstated | Cost kind unknown |
15[7–23] OpenAIDossier | No effort set | Codelia | GPT-5.3 Codex, Accuracy: 75.7% (scale 0.0 to 100.0), 95% confidence interval 71.5 to 80.075.7%(71.5–80.0) | Runs unstated | Cost kind unknown |
16[8–25] AnthropicDossier | No effort set | Capy | Opus 4.6, Accuracy: 75.3% (scale 0.0 to 100.0), 95% confidence interval 70.6 to 80.075.3%(70.6–80.0) | Runs unstated | Cost kind unknown |
17[7–25] OpenAIDossier | No effort set | Simple Codex | GPT-5.3 Codex, Accuracy: 75.1% (scale 0.0 to 100.0), 95% confidence interval 70.4 to 79.775.1%(70.4–79.7) | Runs unstated | Cost kind unknown |
18[9–26] GoogleDossier | No effort set | Terminus-KIRA | Gemini 3.1 Pro Preview, Accuracy: 74.8% (scale 0.0 to 100.0), 95% confidence interval 69.8 to 79.974.8%(69.8–79.9) | Runs unstated | Cost kind unknown |
19[8–26] AnthropicDossier | No effort set | Terminus-KIRA | Opus 4.6, Accuracy: 74.7% (scale 0.0 to 100.0), 95% confidence interval 69.6 to 79.874.7%(69.6–79.8) | Runs unstated | Cost kind unknown |
20[8–29] OpenAIDossier | No effort set | Mux | GPT-5.3 Codex, Accuracy: 74.6% (scale 0.0 to 100.0), 95% confidence interval 69.7 to 79.574.6%(69.7–79.5) | Runs unstated | Cost kind unknown |
21[15–29] AnthropicDossier | No effort set | MAYA-V2 | Opus 4.6, Accuracy: 72.1% (scale 0.0 to 100.0), 95% confidence interval 67.7 to 76.472.1%(67.7–76.4) | Runs unstated | Cost kind unknown |
22[13–32] AnthropicDossier | No effort set | TongAgents | Opus 4.6, Accuracy: 71.9% (scale 0.0 to 100.0), 95% confidence interval 66.7 to 77.271.9%(66.7–77.2) | Runs unstated | Cost kind unknown |
- 1OpenAIDossier
NexAU-AHE
Accuracy
GPT-5.5, Accuracy: 84.7% (scale 0.0 to 100.0), 95% confidence interval 80.6 to 88.884.7%(80.6–88.8) Accuracy
GPT-5.5, Accuracy: 83.1% (scale 0.0 to 100.0), 95% confidence interval 79.0 to 87.383.1%(79.0–87.3)Accuracy
GPT-5.5, Accuracy: 82.2% (scale 0.0 to 100.0), 95% confidence interval 77.9 to 86.682.2%(77.9–86.6)Accuracy
GPT-5.5, Accuracy: 82% (scale 0.0 to 100.0), 95% confidence interval 77.7 to 86.382%(77.7–86.3)Accuracy
GPT-5.4, Accuracy: 81.8% (scale 0.0 to 100.0), 95% confidence interval 77.9 to 85.781.8%(77.9–85.7)Accuracy
Gemini 3.1 Pro Preview, Accuracy: 80.2% (scale 0.0 to 100.0), 95% confidence interval 75.1 to 85.480.2%(75.1–85.4)Accuracy
Opus 4.7, Accuracy: 80.2% (scale 0.0 to 100.0), 95% confidence interval 76.1 to 84.480.2%(76.1–84.4)Accuracy
Opus 4.6, Accuracy: 79.8% (scale 0.0 to 100.0), 95% confidence interval 76.7 to 82.979.8%(76.7–82.9)- 9OpenAIDossier
SageAgent
Accuracy
GPT-5.3 Codex, Accuracy: 78.4% (scale 0.0 to 100.0), 95% confidence interval 74.1 to 82.778.4%(74.1–82.7) - 10GoogleDossier
Forge Code
Accuracy
Gemini 3.1 Pro Preview, Accuracy: 78.4% (scale 0.0 to 100.0), 95% confidence interval 74.9 to 81.978.4%(74.9–81.9) - 10GoogleDossier
ForgeCode
Accuracy
Gemini 3.1 Pro Preview, Accuracy: 78.4% (scale 0.0 to 100.0), 95% confidence interval 74.9 to 81.978.4%(74.9–81.9) - 12OpenAIDossier
Droid
Accuracy
GPT-5.3 Codex, Accuracy: 77.3% (scale 0.0 to 100.0), 95% confidence interval 73.1 to 81.577.3%(73.1–81.5) - 13AnthropicDossier
Meta-Harness
Accuracy
Opus 4.6, Accuracy: 76.4% (scale 0.0 to 100.0), 95% confidence interval 71.6 to 81.276.4%(71.6–81.2) - 14OpenAIDossier
CodeBrain-1.5
Accuracy
GPT-5.3 Codex, Accuracy: 75.8% (scale 0.0 to 100.0), 95% confidence interval 71.9 to 79.875.8%(71.9–79.8) - 15OpenAIDossier
Codelia
Accuracy
GPT-5.3 Codex, Accuracy: 75.7% (scale 0.0 to 100.0), 95% confidence interval 71.5 to 80.075.7%(71.5–80.0) - 16AnthropicDossier
Capy
Accuracy
Opus 4.6, Accuracy: 75.3% (scale 0.0 to 100.0), 95% confidence interval 70.6 to 80.075.3%(70.6–80.0) - 17OpenAIDossier
Simple Codex
Accuracy
GPT-5.3 Codex, Accuracy: 75.1% (scale 0.0 to 100.0), 95% confidence interval 70.4 to 79.775.1%(70.4–79.7) - 18GoogleDossier
Terminus-KIRA
Accuracy
Gemini 3.1 Pro Preview, Accuracy: 74.8% (scale 0.0 to 100.0), 95% confidence interval 69.8 to 79.974.8%(69.8–79.9) - 19AnthropicDossier
Terminus-KIRA
Accuracy
Opus 4.6, Accuracy: 74.7% (scale 0.0 to 100.0), 95% confidence interval 69.6 to 79.874.7%(69.6–79.8) - 20OpenAIDossier
Mux
Accuracy
GPT-5.3 Codex, Accuracy: 74.6% (scale 0.0 to 100.0), 95% confidence interval 69.7 to 79.574.6%(69.7–79.5) - 21AnthropicDossier
MAYA-V2
Accuracy
Opus 4.6, Accuracy: 72.1% (scale 0.0 to 100.0), 95% confidence interval 67.7 to 76.472.1%(67.7–76.4) - 22AnthropicDossier
TongAgents
Accuracy
Opus 4.6, Accuracy: 71.9% (scale 0.0 to 100.0), 95% confidence interval 66.7 to 77.271.9%(66.7–77.2) - 23OpenAIDossier
spoox-o-m
Accuracy
GPT-5.3 Codex, Accuracy: 71.5% (scale 0.0 to 100.0), 95% confidence interval 66.6 to 76.371.5%(66.6–76.3) - 24OpenAIDossier
CodeBrain-1
Accuracy
GPT-5.3 Codex, Accuracy: 70.3% (scale 0.0 to 100.0), 95% confidence interval 65.2 to 75.470.3%(65.2–75.4) - 25AnthropicDossier
Droid
Accuracy
Opus 4.6, Accuracy: 69.9% (scale 0.0 to 100.0), 95% confidence interval 65.0 to 74.869.9%(65.0–74.8) - 26AnthropicDossier
Droid
Accuracy
Opus 4.6, Accuracy: 69.9% (scale 0.0 to 100.0), 95% confidence interval 65.0 to 74.769.9%(65.0–74.7) - 27GoogleDossier
Ante
Accuracy
Gemini 3 Pro Preview, Accuracy: 69.4% (scale 0.0 to 100.0), 95% confidence interval 65.3 to 73.569.4%(65.3–73.5) - 28OpenAIDossier
IndusAGI Coding Agent
Accuracy
GPT-5.3 Codex, Accuracy: 69.1% (scale 0.0 to 100.0), 95% confidence interval 64.7 to 73.669.1%(64.7–73.6) - 29AnthropicDossier
Crux
Accuracy
Opus 4.6, Accuracy: 66.9% (scale 0.0 to 100.0)66.9% - 30OpenAIDossier
Deep Agents
Accuracy
GPT-5.2 Codex, Accuracy: 66.5% (scale 0.0 to 100.0), 95% confidence interval 60.5 to 72.566.5%(60.5–72.5) - 30AnthropicDossier
Mux
Accuracy
Opus 4.6, Accuracy: 66.5% (scale 0.0 to 100.0), 95% confidence interval 61.6 to 71.566.5%(61.6–71.5) - 32OpenAIDossier
clnkr
Accuracy
GPT-5.5, Accuracy: 66.1% (scale 0.0 to 100.0), 95% confidence interval 61.2 to 71.066.1%(61.2–71.0) - 33GoogleDossier
SageAgent
Accuracy
Gemini 3 Pro Preview, Accuracy: 65.2% (scale 0.0 to 100.0), 95% confidence interval 61.1 to 69.265.2%(61.1–69.2) - 34OpenAIDossier
Droid
Accuracy
GPT-5.2, Accuracy: 64.9% (scale 0.0 to 100.0), 95% confidence interval 59.5 to 70.464.9%(59.5–70.4) - 35OpenAIDossier
medium · Droid
Accuracy
GPT-5.2, Accuracy: 64.9% (scale 0.0 to 100.0), 95% confidence interval 59.4 to 70.464.9%(59.4–70.4) - 36OpenAIDossier
Terminus 2
Accuracy
GPT-5.3 Codex, Accuracy: 64.7% (scale 0.0 to 100.0), 95% confidence interval 59.4 to 70.064.7%(59.4–70.0) - 37GoogleDossier
Junie CLI
Accuracy
Gemini 3 Flash Preview, Accuracy: 64.3% (scale 0.0 to 100.0), 95% confidence interval 58.8 to 69.764.3%(58.8–69.7) - 38AnthropicDossier
Droid
Accuracy
Opus 4.5, Accuracy: 63.1% (scale 0.0 to 100.0), 95% confidence interval 57.9 to 68.463.1%(57.9–68.4) - 39AnthropicDossier
Droid
Accuracy
Opus 4.5, Accuracy: 63.1% (scale 0.0 to 100.0), 95% confidence interval 57.8 to 68.463.1%(57.8–68.4) - 40OpenAIDossier
Codex CLI
Accuracy
GPT-5.2, Accuracy: 62.9% (scale 0.0 to 100.0), 95% confidence interval 57.0 to 68.862.9%(57.0–68.8) - 40AnthropicDossier
Terminus 2
Accuracy
Opus 4.6, Accuracy: 62.9% (scale 0.0 to 100.0), 95% confidence interval 57.7 to 68.162.9%(57.7–68.1) - 42OpenAIDossier
medium · Codex CLI
Accuracy
GPT-5.2, Accuracy: 62.9% (scale 0.0 to 100.0), 95% confidence interval 57.0 to 68.862.9%(57.0–68.8) - 43GoogleDossier
CodeBrain-1.5
Accuracy
Gemini 3 Pro Preview, Accuracy: 62.2% (scale 0.0 to 100.0), 95% confidence interval 57.1 to 67.462.2%(57.1–67.4) - 44GoogleDossier
CodeBrain-1
Accuracy
Gemini 3 Pro Preview, Accuracy: 62.2% (scale 0.0 to 100.0), 95% confidence interval 57.1 to 67.362.2%(57.1–67.3) - 45GoogleDossier
II-Agent
Accuracy
Gemini 3 Pro Preview, Accuracy: 61.8% (scale 0.0 to 100.0), 95% confidence interval 56.3 to 67.361.8%(56.3–67.3) - 46OpenAIDossier
hookele
Accuracy
GPT-5.1 Codex mini, Accuracy: 61.6% (scale 0.0 to 100.0), 95% confidence interval 57.9 to 65.361.6%(57.9–65.3) - 47GoogleDossier
Gemini CLI
Accuracy
Gemini 3.1 Pro Preview, Accuracy: 61.4% (scale 0.0 to 100.0), 95% confidence interval 53.3 to 69.561.4%(53.3–69.5) - 48GoogleDossier
Droid
Accuracy
Gemini 3 Pro Preview, Accuracy: 61.1% (scale 0.0 to 100.0), 95% confidence interval 55.7 to 66.661.1%(55.7–66.6) - 49OpenAIDossier
Mux
Accuracy
GPT-5.2, Accuracy: 60.7% (scale 0.0 to 100.0)60.7% - 50OpenAIDossier
Codex CLI
Accuracy
GPT-5.1 Codex Max, Accuracy: 60.4% (scale 0.0 to 100.0), 95% confidence interval 55.2 to 65.760.4%(55.2–65.7) - 51AnthropicDossier
Letta Code
Accuracy
Opus 4.5, Accuracy: 59.1% (scale 0.0 to 100.0), 95% confidence interval 54.4 to 63.859.1%(54.4–63.8) - 52AnthropicDossier
128,000 budget · Letta Code
Accuracy
Opus 4.5, Accuracy: 59.1% (scale 0.0 to 100.0), 95% confidence interval 54.4 to 63.859.1%(54.4–63.8) - 53AnthropicDossier
Mux
Accuracy
Opus 4.5, Accuracy: 58.4% (scale 0.0 to 100.0)58.4% - 54AnthropicDossier
Claude Code
Accuracy
Opus 4.6, Accuracy: 58% (scale 0.0 to 100.0), 95% confidence interval 52.3 to 63.758%(52.3–63.7) - 55OpenAIDossier
Codex CLI
Accuracy
GPT-5.1 Codex, Accuracy: 57.8% (scale 0.0 to 100.0), 95% confidence interval 52.1 to 63.557.8%(52.1–63.5) - 56OpenAIDossier
Crux
Accuracy
GPT-5.1 Codex, Accuracy: 57.8% (scale 0.0 to 100.0), 95% confidence interval 52.0 to 63.557.8%(52.0–63.5) - 56AnthropicDossier
Terminus 2
Accuracy
Opus 4.5, Accuracy: 57.8% (scale 0.0 to 100.0), 95% confidence interval 52.8 to 62.757.8%(52.8–62.7) - 58xAIDossier
Grok CLI
Accuracy
Grok 4.20, Accuracy: 57.3% (scale 0.0 to 100.0)57.3% - 59xAIDossier
grok-cli
Accuracy
Grok 4.20, Accuracy: 57.3% (scale 0.0 to 100.0)57.3% - 60GoogleDossier
Terminus 2
Accuracy
Gemini 3 Pro Preview, Accuracy: 56.9% (scale 0.0 to 100.0), 95% confidence interval 52.0 to 61.756.9%(52.0–61.7) - 61GoogleDossier
Letta Code
Accuracy
Gemini 3 Pro Preview, Accuracy: 56% (scale 0.0 to 100.0), 95% confidence interval 50.1 to 61.856%(50.1–61.8) - 62AnthropicDossier
Goose
Accuracy
Opus 4.5, Accuracy: 54.3% (scale 0.0 to 100.0), 95% confidence interval 49.2 to 59.554.3%(49.2–59.5) - 63OpenAIDossier
Terminus 2
Accuracy
GPT-5.2, Accuracy: 54% (scale 0.0 to 100.0), 95% confidence interval 48.4 to 59.754%(48.4–59.7) - 64OpenAIDossier
Letta Code
Accuracy
GPT-5.1 Codex, Accuracy: 53.5% (scale 0.0 to 100.0), 95% confidence interval 47.9 to 59.053.5%(47.9–59.0) - 65AnthropicDossier
Simplai Agent
Accuracy
Sonnet 4.6, Accuracy: 53.4% (scale 0.0 to 100.0), 95% confidence interval 47.9 to 59.053.4%(47.9–59.0) - 66AnthropicDossier
Simplai Agent
Accuracy
Sonnet 4.6, Accuracy: 53.4% (scale 0.0 to 100.0), 95% confidence interval 47.9 to 58.953.4%(47.9–58.9) - 67Z.aiDossier
Terminus 2
Accuracy
GLM-5, Accuracy: 52.4% (scale 0.0 to 100.0), 95% confidence interval 47.3 to 57.652.4%(47.3–57.6) - 68AnthropicDossier
Claude Code
Accuracy
Opus 4.5, Accuracy: 52.1% (scale 0.0 to 100.0), 95% confidence interval 47.2 to 57.152.1%(47.2–57.1) - 69AnthropicDossier
OpenHands
Accuracy
Opus 4.5, Accuracy: 51.9% (scale 0.0 to 100.0), 95% confidence interval 46.2 to 57.651.9%(46.2–57.6) - 70GoogleDossier
Terminus 2
Accuracy
Gemini 3 Flash Preview, Accuracy: 51.7% (scale 0.0 to 100.0), 95% confidence interval 45.6 to 57.851.7%(45.6–57.8) - 70AnthropicDossier
OpenCode
Accuracy
Opus 4.5, Accuracy: 51.7% (scale 0.0 to 100.0)51.7% - 72GoogleDossier
Gemini CLI
Accuracy
Gemini 3 Flash Preview, Accuracy: 51% (scale 0.0 to 100.0), 95% confidence interval 45.1 to 56.951%(45.1–56.9) - 73OpenAIDossier
Codex CLI
Accuracy
GPT-5, Accuracy: 49.6% (scale 0.0 to 100.0), 95% confidence interval 43.9 to 55.349.6%(43.9–55.3) - 74OpenAIDossier
medium · Codex CLI
Accuracy
GPT-5, Accuracy: 49.6% (scale 0.0 to 100.0), 95% confidence interval 43.9 to 55.349.6%(43.9–55.3) - 75OpenAIDossier
medium · Terminus 2
Accuracy
GPT-5.1, Accuracy: 47.6% (scale 0.0 to 100.0), 95% confidence interval 42.1 to 53.147.6%(42.1–53.1) - 75OpenAIDossier
medium · Terminus 2
Accuracy
GPT-5.1, Accuracy: 47.6% (scale 0.0 to 100.0), 95% confidence interval 42.1 to 53.147.6%(42.1–53.1) - 77OpenAIDossier
Terminus 2
Accuracy
GPT-5.1, Accuracy: 47.6% (scale 0.0 to 100.0), 95% confidence interval 42.2 to 53.047.6%(42.2–53.0) - 78GoogleDossier
Gemini CLI
Accuracy
Gemini 3 Flash Preview, Accuracy: 47.4% (scale 0.0 to 100.0), 95% confidence interval 41.5 to 53.447.4%(41.5–53.4) - 79AnthropicDossier
CAMEL-AI
Accuracy
Sonnet 4.5, Accuracy: 46.5% (scale 0.0 to 100.0), 95% confidence interval 41.9 to 51.146.5%(41.9–51.1) - 80MiniMaxDossier
IndusAGI Coding Agent
Accuracy
MiniMax M2.7, Accuracy: 45.1% (scale 0.0 to 100.0)45.1% - 81OpenAIDossier
Codex CLI
Accuracy
GPT-5 Codex, Accuracy: 44.3% (scale 0.0 to 100.0), 95% confidence interval 39.0 to 49.644.3%(39.0–49.6) - 82OpenAIDossier
Codex CLI
Accuracy
GPT-5 Codex, Accuracy: 44.3% (scale 0.0 to 100.0), 95% confidence interval 39.0 to 49.644.3%(39.0–49.6) - 83OpenAIDossier
OpenHands
Accuracy
GPT-5, Accuracy: 43.8% (scale 0.0 to 100.0), 95% confidence interval 38.0 to 49.743.8%(38.0–49.7) - 84OpenAIDossier
medium · OpenHands
Accuracy
GPT-5, Accuracy: 43.8% (scale 0.0 to 100.0), 95% confidence interval 37.9 to 49.743.8%(37.9–49.7) - 85OpenAIDossier
Terminus 2
Accuracy
GPT-5 Codex, Accuracy: 43.4% (scale 0.0 to 100.0), 95% confidence interval 37.8 to 49.143.4%(37.8–49.1) - 86OpenAIDossier
Terminus 2
Accuracy
GPT-5 Codex, Accuracy: 43.4% (scale 0.0 to 100.0), 95% confidence interval 37.7 to 49.143.4%(37.7–49.1) - 87Moonshot AIDossier
Terminus 2
Accuracy
Kimi K2.5, Accuracy: 43.2% (scale 0.0 to 100.0), 95% confidence interval 37.6 to 48.843.2%(37.6–48.8) - 88OpenAIDossier
Crux
Accuracy
GPT-5.1 Codex mini, Accuracy: 43.1% (scale 0.0 to 100.0), 95% confidence interval 37.3 to 49.043.1%(37.3–49.0) - 88AnthropicDossier
Goose
Accuracy
Sonnet 4.5, Accuracy: 43.1% (scale 0.0 to 100.0), 95% confidence interval 38.1 to 48.243.1%(38.1–48.2) - 90OpenAIDossier
Codex CLI
Accuracy
GPT-5.1 Codex mini, Accuracy: 43.1% (scale 0.0 to 100.0), 95% confidence interval 37.2 to 49.043.1%(37.2–49.0) - 91MiniMaxDossier
Harness Agent
Accuracy
MiniMax M2.7, Accuracy: 42.9% (scale 0.0 to 100.0), 95% confidence interval 37.2 to 48.742.9%(37.2–48.7) - 92AnthropicDossier
Terminus 2
Accuracy
Sonnet 4.5, Accuracy: 42.8% (scale 0.0 to 100.0), 95% confidence interval 37.3 to 48.342.8%(37.3–48.3) - 93AnthropicDossier
Terminus 2
Accuracy
Sonnet 4.5, Accuracy: 42.8% (scale 0.0 to 100.0), 95% confidence interval 37.3 to 48.342.8%(37.3–48.3) - 94AnthropicDossier
MAYA
Accuracy
Sonnet 4.5, Accuracy: 42.7% (scale 0.0 to 100.0)42.7% - 95MiniMaxDossier
cchuter
Accuracy
MiniMax M2.5, Accuracy: 42.7% (scale 0.0 to 100.0), 95% confidence interval 37.2 to 48.242.7%(37.2–48.2) - 95AnthropicDossier
MAYA-V2
Accuracy
Sonnet 4.5, Accuracy: 42.7% (scale 0.0 to 100.0)42.7% - 97AnthropicDossier
OpenHands
Accuracy
Sonnet 4.5, Accuracy: 42.6% (scale 0.0 to 100.0), 95% confidence interval 37.1 to 48.142.6%(37.1–48.1) - 98AnthropicDossier
OpenHands
Accuracy
Sonnet 4.5, Accuracy: 42.6% (scale 0.0 to 100.0), 95% confidence interval 37.2 to 48.042.6%(37.2–48.0) - 99AnthropicDossier
Mini-SWE-Agent
Accuracy
Sonnet 4.5, Accuracy: 42.5% (scale 0.0 to 100.0), 95% confidence interval 37.0 to 48.042.5%(37.0–48.0) - 100AnthropicDossier
Mini-SWE-Agent
Accuracy
Sonnet 4.5, Accuracy: 42.5% (scale 0.0 to 100.0), 95% confidence interval 37.0 to 48.042.5%(37.0–48.0) - 101MiniMaxDossier
Terminus 2
Accuracy
MiniMax M2.5, Accuracy: 42.2% (scale 0.0 to 100.0), 95% confidence interval 37.1 to 47.442.2%(37.1–47.4) - 102OpenAIDossier
Mini-SWE-Agent
Accuracy
GPT-5 Codex, Accuracy: 41.3% (scale 0.0 to 100.0), 95% confidence interval 35.9 to 46.841.3%(35.9–46.8) - 103OpenAIDossier
Mini-SWE-Agent
Accuracy
GPT-5 Codex, Accuracy: 41.3% (scale 0.0 to 100.0), 95% confidence interval 35.8 to 46.841.3%(35.8–46.8) - 104AnthropicDossier
Claude Code
Accuracy
Sonnet 4.5, Accuracy: 40.1% (scale 0.0 to 100.0), 95% confidence interval 34.4 to 45.840.1%(34.4–45.8) - 105AnthropicDossier
Claude Code
Accuracy
Sonnet 4.5, Accuracy: 40.1% (scale 0.0 to 100.0), 95% confidence interval 34.3 to 45.840.1%(34.3–45.8) - 106DeepSeekDossier
Terminus 2
Accuracy
DeepSeek V3.2, Accuracy: 39.6% (scale 0.0 to 100.0), 95% confidence interval 34.2 to 44.939.6%(34.2–44.9) - 107AnthropicDossier
Terminus 2
Accuracy
Opus 4.1, Accuracy: 38% (scale 0.0 to 100.0), 95% confidence interval 32.9 to 43.138%(32.9–43.1) - 108AnthropicDossier
Terminus 2
Accuracy
Opus 4.1, Accuracy: 38% (scale 0.0 to 100.0), 95% confidence interval 32.8 to 43.138%(32.8–43.1) - 109OpenAIDossier
Terminus 2
Accuracy
GPT-5.1 Codex, Accuracy: 36.9% (scale 0.0 to 100.0), 95% confidence interval 30.6 to 43.236.9%(30.6–43.2) - 109AnthropicDossier
OpenHands
Accuracy
Opus 4.1, Accuracy: 36.9% (scale 0.0 to 100.0), 95% confidence interval 31.6 to 42.236.9%(31.6–42.2) - 111OpenAIDossier
Terminus 2
Accuracy
GPT-5.1 Codex, Accuracy: 36.9% (scale 0.0 to 100.0), 95% confidence interval 30.5 to 43.236.9%(30.5–43.2) - 111AnthropicDossier
OpenHands
Accuracy
Opus 4.1, Accuracy: 36.9% (scale 0.0 to 100.0), 95% confidence interval 31.6 to 42.136.9%(31.6–42.1) - 113MiniMaxDossier
Crux
Accuracy
MiniMax M2.1, Accuracy: 36.6% (scale 0.0 to 100.0), 95% confidence interval 31.0 to 42.336.6%(31.0–42.3) - 114Moonshot AIDossier
Terminus 2
Accuracy
Kimi K2 Thinking, Accuracy: 35.7% (scale 0.0 to 100.0), 95% confidence interval 30.2 to 41.335.7%(30.2–41.3) - 115Moonshot AIDossier
Terminus 2
Accuracy
Kimi K2 Thinking, Accuracy: 35.7% (scale 0.0 to 100.0), 95% confidence interval 30.2 to 41.235.7%(30.2–41.2) - 116AnthropicDossier
Goose
Accuracy
Haiku 4.5, Accuracy: 35.5% (scale 0.0 to 100.0), 95% confidence interval 29.8 to 41.235.5%(29.8–41.2) - 117OpenAIDossier
medium · Terminus 2
Accuracy
GPT-5, Accuracy: 35.2% (scale 0.0 to 100.0), 95% confidence interval 29.1 to 41.335.2%(29.1–41.3) - 118OpenAIDossier
Terminus 2
Accuracy
GPT-5, Accuracy: 35.2% (scale 0.0 to 100.0), 95% confidence interval 29.2 to 41.235.2%(29.2–41.2) - 119AnthropicDossier
Mini-SWE-Agent
Accuracy
Opus 4.1, Accuracy: 35.1% (scale 0.0 to 100.0), 95% confidence interval 30.2 to 40.035.1%(30.2–40.0) - 120AnthropicDossier
Mini-SWE-Agent
Accuracy
Opus 4.1, Accuracy: 35.1% (scale 0.0 to 100.0), 95% confidence interval 30.1 to 40.035.1%(30.1–40.0) - 121OpenAIDossier
spoox-o-m
Accuracy
GPT-5 mini, Accuracy: 34.8% (scale 0.0 to 100.0), 95% confidence interval 29.5 to 40.234.8%(29.5–40.2) - 121AnthropicDossier
Claude Code
Accuracy
Opus 4.1, Accuracy: 34.8% (scale 0.0 to 100.0), 95% confidence interval 29.1 to 40.534.8%(29.1–40.5) - 123OpenAIDossier
spoox-m
Accuracy
GPT-5 mini, Accuracy: 34.8% (scale 0.0 to 100.0), 95% confidence interval 29.5 to 40.134.8%(29.5–40.1) - 123AnthropicDossier
Claude Code
Accuracy
Opus 4.1, Accuracy: 34.8% (scale 0.0 to 100.0), 95% confidence interval 29.1 to 40.534.8%(29.1–40.5) - 125OpenAIDossier
Mini-SWE-Agent
Accuracy
GPT-5, Accuracy: 33.9% (scale 0.0 to 100.0), 95% confidence interval 28.2 to 39.633.9%(28.2–39.6) - 126OpenAIDossier
medium · Mini-SWE-Agent
Accuracy
GPT-5, Accuracy: 33.9% (scale 0.0 to 100.0), 95% confidence interval 28.2 to 39.633.9%(28.2–39.6) - 127Z.aiDossier
Terminus 2
Accuracy
GLM-4.7, Accuracy: 33.4% (scale 0.0 to 100.0), 95% confidence interval 27.8 to 39.033.4%(27.8–39.0) - 128Z.aiDossier
Crux
Accuracy
GLM-4.7, Accuracy: 33.3% (scale 0.0 to 100.0), 95% confidence interval 28.5 to 38.133.3%(28.5–38.1) - 129GoogleDossier
Terminus 2
Accuracy
Gemini 2.5 Pro, Accuracy: 32.6% (scale 0.0 to 100.0), 95% confidence interval 26.7 to 38.632.6%(26.7–38.6) - 130GoogleDossier
Terminus 2
Accuracy
Gemini 2.5 Pro, Accuracy: 32.6% (scale 0.0 to 100.0), 95% confidence interval 26.7 to 38.532.6%(26.7–38.5) - 131OpenAIDossier
medium · Codex CLI
Accuracy
GPT-5 mini, Accuracy: 31.9% (scale 0.0 to 100.0), 95% confidence interval 26.0 to 37.831.9%(26.0–37.8) - 132OpenAIDossier
Codex CLI
Accuracy
GPT-5 mini, Accuracy: 31.9% (scale 0.0 to 100.0), 95% confidence interval 26.0 to 37.731.9%(26.0–37.7) - 133MiniMaxDossier
Terminus 2
Accuracy
MiniMax M2, Accuracy: 30% (scale 0.0 to 100.0), 95% confidence interval 24.7 to 35.430%(24.7–35.4) - 134AnthropicDossier
Mini-SWE-Agent
Accuracy
Haiku 4.5, Accuracy: 29.8% (scale 0.0 to 100.0), 95% confidence interval 24.9 to 34.829.8%(24.9–34.8) - 135AnthropicDossier
Mini-SWE-Agent
Accuracy
Haiku 4.5, Accuracy: 29.8% (scale 0.0 to 100.0), 95% confidence interval 24.9 to 34.729.8%(24.9–34.7) - 136OpenAIDossier
OpenHands
Accuracy
GPT-5 mini, Accuracy: 29.2% (scale 0.0 to 100.0), 95% confidence interval 23.8 to 34.629.2%(23.8–34.6) - 136MiniMaxDossier
Terminus 2
Accuracy
MiniMax M2.1, Accuracy: 29.2% (scale 0.0 to 100.0), 95% confidence interval 23.5 to 34.929.2%(23.5–34.9) - 138OpenAIDossier
medium · OpenHands
Accuracy
GPT-5 mini, Accuracy: 29.2% (scale 0.0 to 100.0), 95% confidence interval 23.7 to 34.729.2%(23.7–34.7) - 139AnthropicDossier
Terminus 2
Accuracy
Haiku 4.5, Accuracy: 28.3% (scale 0.0 to 100.0), 95% confidence interval 22.7 to 34.028.3%(22.7–34.0) - 140AnthropicDossier
Terminus 2
Accuracy
Haiku 4.5, Accuracy: 28.3% (scale 0.0 to 100.0), 95% confidence interval 22.6 to 34.028.3%(22.6–34.0) - 141Moonshot AIDossier
Terminus 2
Accuracy
Kimi K2, Accuracy: 27.8% (scale 0.0 to 100.0), 95% confidence interval 22.9 to 32.727.8%(22.9–32.7) - 142Moonshot AIDossier
Terminus 2
Accuracy
Kimi K2, Accuracy: 27.8% (scale 0.0 to 100.0), 95% confidence interval 22.8 to 32.727.8%(22.8–32.7) - 143AnthropicDossier
Claude Code
Accuracy
Haiku 4.5, Accuracy: 27.5% (scale 0.0 to 100.0), 95% confidence interval 22.1 to 33.027.5%(22.1–33.0) - 144AnthropicDossier
Claude Code
Accuracy
Haiku 4.5, Accuracy: 27.5% (scale 0.0 to 100.0), 95% confidence interval 22.0 to 33.027.5%(22.0–33.0) - 145xAIDossier
OpenHands
Accuracy
Grok 4 (0709), Accuracy: 27.2% (scale 0.0 to 100.0), 95% confidence interval 21.1 to 33.327.2%(21.1–33.3) - 146xAIDossier
OpenHands
Accuracy
Grok 4 (0709), Accuracy: 27.2% (scale 0.0 to 100.0), 95% confidence interval 21.2 to 33.227.2%(21.2–33.2) - 146AlibabaDossier
Dakou Agent
Accuracy
Qwen3 Coder 480B A35B Instruct, Accuracy: 27.2% (scale 0.0 to 100.0), 95% confidence interval 22.2 to 32.227.2%(22.2–32.2) - 148Moonshot AIDossier
OpenHands
Accuracy
Kimi K2, Accuracy: 26.7% (scale 0.0 to 100.0), 95% confidence interval 21.5 to 32.026.7%(21.5–32.0) - 149Moonshot AIDossier
OpenHands
Accuracy
Kimi K2, Accuracy: 26.7% (scale 0.0 to 100.0), 95% confidence interval 21.4 to 32.026.7%(21.4–32.0) - 150GoogleDossier
Mini-SWE-Agent
Accuracy
Gemini 2.5 Pro, Accuracy: 26.1% (scale 0.0 to 100.0), 95% confidence interval 21.2 to 31.026.1%(21.2–31.0) - 151GoogleDossier
Mini-SWE-Agent
Accuracy
Gemini 2.5 Pro, Accuracy: 26.1% (scale 0.0 to 100.0), 95% confidence interval 21.1 to 31.126.1%(21.1–31.1) - 152xAIDossier
Mini-SWE-Agent
Accuracy
Grok Code Fast 1, Accuracy: 25.8% (scale 0.0 to 100.0), 95% confidence interval 20.7 to 31.025.8%(20.7–31.0) - 153xAIDossier
Mini-SWE-Agent
Accuracy
Grok Code Fast 1, Accuracy: 25.8% (scale 0.0 to 100.0), 95% confidence interval 20.7 to 30.925.8%(20.7–30.9) - 154xAIDossier
Mini-SWE-Agent
Accuracy
Grok 4 (0709), Accuracy: 25.4% (scale 0.0 to 100.0), 95% confidence interval 19.7 to 31.125.4%(19.7–31.1) - 154AlibabaDossier
OpenHands
Accuracy
Qwen3 Coder 480B A35B Instruct, Accuracy: 25.4% (scale 0.0 to 100.0), 95% confidence interval 20.3 to 30.525.4%(20.3–30.5) - 156xAIDossier
Mini-SWE-Agent
Accuracy
Grok 4 (0709), Accuracy: 25.4% (scale 0.0 to 100.0), 95% confidence interval 19.7 to 31.125.4%(19.7–31.1) - 156AlibabaDossier
OpenHands
Accuracy
Qwen3 Coder 480B A35B Instruct, Accuracy: 25.4% (scale 0.0 to 100.0), 95% confidence interval 20.3 to 30.525.4%(20.3–30.5) - 158AlibabaDossier
little-coder
Accuracy
Qwen3.6 35B A3B, Accuracy: 24.6% (scale 0.0 to 100.0), 95% confidence interval 18.3 to 30.924.6%(18.3–30.9) - 159Z.aiDossier
Terminus 2
Accuracy
GLM-4.6, Accuracy: 24.5% (scale 0.0 to 100.0), 95% confidence interval 19.7 to 29.324.5%(19.7–29.3) - 160Z.aiDossier
Terminus 2
Accuracy
GLM-4.6, Accuracy: 24.5% (scale 0.0 to 100.0), 95% confidence interval 19.8 to 29.224.5%(19.8–29.2) - 161OpenAIDossier
Terminus 2
Accuracy
GPT-5 mini, Accuracy: 24% (scale 0.0 to 100.0), 95% confidence interval 19.1 to 29.024%(19.1–29.0) - 162OpenAIDossier
medium · Terminus 2
Accuracy
GPT-5 mini, Accuracy: 24% (scale 0.0 to 100.0), 95% confidence interval 19.1 to 28.924%(19.1–28.9) - 163AlibabaDossier
Terminus 2
Accuracy
Qwen3 Coder 480B A35B Instruct, Accuracy: 23.9% (scale 0.0 to 100.0), 95% confidence interval 18.4 to 29.423.9%(18.4–29.4) - 164AlibabaDossier
Terminus 2
Accuracy
Qwen3 Coder 480B A35B Instruct, Accuracy: 23.9% (scale 0.0 to 100.0), 95% confidence interval 18.4 to 29.423.9%(18.4–29.4) - 165xAIDossier
Terminus 2
Accuracy
Grok 4 (0709), Accuracy: 23.1% (scale 0.0 to 100.0), 95% confidence interval 17.5 to 28.823.1%(17.5–28.8) - 166xAIDossier
Terminus 2
Accuracy
Grok 4 (0709), Accuracy: 23.1% (scale 0.0 to 100.0), 95% confidence interval 17.4 to 28.823.1%(17.4–28.8) - 167OpenAIDossier
Mini-SWE-Agent
Accuracy
GPT-5 mini, Accuracy: 22.2% (scale 0.0 to 100.0), 95% confidence interval 17.1 to 27.422.2%(17.1–27.4) - 168OpenAIDossier
medium · Mini-SWE-Agent
Accuracy
GPT-5 mini, Accuracy: 22.2% (scale 0.0 to 100.0), 95% confidence interval 17.1 to 27.322.2%(17.1–27.3) - 169OpenAIDossier
spoox-o-m
Accuracy
GPT-5 nano, Accuracy: 21.8% (scale 0.0 to 100.0), 95% confidence interval 16.3 to 27.321.8%(16.3–27.3) - 170GoogleDossier
Gemini CLI
Accuracy
Gemini 2.5 Pro, Accuracy: 19.6% (scale 0.0 to 100.0), 95% confidence interval 13.9 to 25.319.6%(13.9–25.3) - 171GoogleDossier
Gemini CLI
Accuracy
Gemini 2.5 Pro, Accuracy: 19.6% (scale 0.0 to 100.0), 95% confidence interval 13.9 to 25.219.6%(13.9–25.2) - 172OpenAIDossier
Terminus 2
Accuracy
gpt-oss-120b, Accuracy: 18.7% (scale 0.0 to 100.0), 95% confidence interval 13.4 to 24.018.7%(13.4–24.0) - 173OpenAIDossier
Terminus 2
Accuracy
gpt-oss-120b, Accuracy: 18.7% (scale 0.0 to 100.0), 95% confidence interval 13.5 to 23.918.7%(13.5–23.9) - 174GoogleDossier
Mini-SWE-Agent
Accuracy
Gemini 2.5 Flash Preview (09-2025), Accuracy: 17.1% (scale 0.0 to 100.0), 95% confidence interval 12.2 to 22.017.1%(12.2–22.0) - 175GoogleDossier
Mini-SWE-Agent
Accuracy
Gemini 2.5 Flash, Accuracy: 17.1% (scale 0.0 to 100.0), 95% confidence interval 12.2 to 22.017.1%(12.2–22.0) - 176GoogleDossier
Terminus 2
Accuracy
Gemini 2.5 Flash Preview (09-2025), Accuracy: 16.9% (scale 0.0 to 100.0), 95% confidence interval 12.2 to 21.616.9%(12.2–21.6) - 177GoogleDossier
Terminus 2
Accuracy
Gemini 2.5 Flash, Accuracy: 16.9% (scale 0.0 to 100.0), 95% confidence interval 12.2 to 21.516.9%(12.2–21.5) - 178GoogleDossier
OpenHands
Accuracy
Gemini 2.5 Flash, Accuracy: 16.4% (scale 0.0 to 100.0), 95% confidence interval 11.7 to 21.116.4%(11.7–21.1) - 178GoogleDossier
OpenHands
Accuracy
Gemini 2.5 Pro, Accuracy: 16.4% (scale 0.0 to 100.0), 95% confidence interval 11.0 to 21.816.4%(11.0–21.8) - 180GoogleDossier
OpenHands
Accuracy
Gemini 2.5 Flash Preview (09-2025), Accuracy: 16.4% (scale 0.0 to 100.0), 95% confidence interval 11.7 to 21.116.4%(11.7–21.1) - 180GoogleDossier
OpenHands
Accuracy
Gemini 2.5 Pro, Accuracy: 16.4% (scale 0.0 to 100.0), 95% confidence interval 10.9 to 21.916.4%(10.9–21.9) - 182GoogleDossier
Gemini CLI
Accuracy
Gemini 2.5 Flash, Accuracy: 15.4% (scale 0.0 to 100.0), 95% confidence interval 10.9 to 20.015.4%(10.9–20.0) - 183GoogleDossier
Gemini CLI
Accuracy
Gemini 2.5 Flash Preview (09-2025), Accuracy: 15.4% (scale 0.0 to 100.0), 95% confidence interval 10.9 to 19.915.4%(10.9–19.9) - 184OpenAIDossier
Mini-SWE-Agent
Accuracy
gpt-oss-120b, Accuracy: 14.2% (scale 0.0 to 100.0), 95% confidence interval 9.7 to 18.714.2%(9.7–18.7) - 184xAIDossier
Terminus 2
Accuracy
Grok Code Fast 1, Accuracy: 14.2% (scale 0.0 to 100.0), 95% confidence interval 9.3 to 19.114.2%(9.3–19.1) - 186OpenAIDossier
Mini-SWE-Agent
Accuracy
gpt-oss-120b, Accuracy: 14.2% (scale 0.0 to 100.0), 95% confidence interval 9.7 to 18.614.2%(9.7–18.6) - 186xAIDossier
Terminus 2
Accuracy
Grok Code Fast 1, Accuracy: 14.2% (scale 0.0 to 100.0), 95% confidence interval 9.2 to 19.214.2%(9.2–19.2) - 188AnthropicDossier
OpenHands
Accuracy
Haiku 4.5, Accuracy: 13.9% (scale 0.0 to 100.0), 95% confidence interval 8.6 to 19.313.9%(8.6–19.3) - 189AnthropicDossier
OpenHands
Accuracy
Haiku 4.5, Accuracy: 13.9% (scale 0.0 to 100.0), 95% confidence interval 8.6 to 19.213.9%(8.6–19.2) - 190OpenAIDossier
medium · Codex CLI
Accuracy
GPT-5 nano, Accuracy: 11.5% (scale 0.0 to 100.0), 95% confidence interval 7.0 to 16.011.5%(7.0–16.0) - 191OpenAIDossier
Codex CLI
Accuracy
GPT-5 nano, Accuracy: 11.5% (scale 0.0 to 100.0), 95% confidence interval 7.0 to 15.911.5%(7.0–15.9) - 192OpenAIDossier
medium · OpenHands
Accuracy
GPT-5 nano, Accuracy: 9.9% (scale 0.0 to 100.0), 95% confidence interval 5.8 to 14.09.9%(5.8–14.0) - 193OpenAIDossier
OpenHands
Accuracy
GPT-5 nano, Accuracy: 9.9% (scale 0.0 to 100.0), 95% confidence interval 5.7 to 14.09.9%(5.7–14.0) - 194AlibabaDossier
little-coder
Accuracy
Qwen3.5 9B, Accuracy: 9.2% (scale 0.0 to 100.0), 95% confidence interval 4.6 to 13.89.2%(4.6–13.8) - 195OpenAIDossier
medium · Terminus 2
Accuracy
GPT-5 nano, Accuracy: 7.9% (scale 0.0 to 100.0), 95% confidence interval 4.2 to 11.67.9%(4.2–11.6) - 196OpenAIDossier
Terminus 2
Accuracy
GPT-5 nano, Accuracy: 7.9% (scale 0.0 to 100.0), 95% confidence interval 4.2 to 11.57.9%(4.2–11.5) - 197OpenAIDossier
medium · Mini-SWE-Agent
Accuracy
GPT-5 nano, Accuracy: 7% (scale 0.0 to 100.0), 95% confidence interval 3.3 to 10.77%(3.3–10.7) - 198OpenAIDossier
Mini-SWE-Agent
Accuracy
GPT-5 nano, Accuracy: 7% (scale 0.0 to 100.0), 95% confidence interval 3.2 to 10.77%(3.2–10.7) - 199OpenAIDossier
Mini-SWE-Agent
Accuracy
gpt-oss-20b, Accuracy: 3.4% (scale 0.0 to 100.0), 95% confidence interval 0.7 to 6.13.4%(0.7–6.1) - 200OpenAIDossier
Mini-SWE-Agent
Accuracy
gpt-oss-20b, Accuracy: 3.4% (scale 0.0 to 100.0), 95% confidence interval 0.7 to 6.03.4%(0.7–6.0) - 201OpenAIDossier
Terminus 2
Accuracy
gpt-oss-20b, Accuracy: 3.1% (scale 0.0 to 100.0), 95% confidence interval 0.2 to 6.03.1%(0.2–6.0) - 202OpenAIDossier
Terminus 2
Accuracy
gpt-oss-20b, Accuracy: 3.1% (scale 0.0 to 100.0), 95% confidence interval 0.1 to 6.03.1%(0.1–6.0)