LLM Benchmarks — Scores, Methodology & Live Rankings
What each benchmark measures, how models score, and the current rankings
Terminal-Bench Lite
Agentic Coding
Lite variant of Terminal-Bench (terminal coding agent evaluations) as reported on the KRAFTON Raon-1 mini preview card; evaluated with judge gpt-5.6-luna. Di...
Terminal-Bench Science 0.1
Agentic Coding
Terminal-Bench Science task variant (leaderboard task id terminalbenchscience_0_1), evaluated by Artificial Analysis on the harborframework/terminal-bench-sc...
FrontierCode 2.0
Agentic Coding
Cognition private benchmark assessing end-to-end code quality and whether a repository maintainer would merge the agent's change. Mistral blog: Mistral Large...
SWE Refactor Bench
Agentic Coding
Agentic software refactoring benchmark; agent harness Claude Code 2.1.220 at 200K context. Introduced on the MiniMax M3.1-Flash release card (M3.1-Flash + fr...
BenchCAD w/o tools
Agentic Coding
BenchCAD on 1,000-case subsets randomly sampled from the 17,900-case Vision2Code dataset, max reasoning effort, single-shot generation with official prompts ...
BenchCAD w/ tools
Agentic Coding
With-tools run uses Claude Code; open models scored on 1,000-case subsets of Vision2Code.
Kimi Code Bench 3.0 Long-horizon
Agentic Coding
Internal Moonshot long-horizon coding benchmark; Kimi tested with Claude Code, competitors at max effort with possible fallback behavior.
SWE-bench Pro v2.0 Hard
Agentic Coding
SWE-bench Pro v2.0 Hard split as reported on the Ripple-65B Preview card; Terminus 2 harness used for Ripple-65B Preview, competitor scores partially officia...
Aider Polyglot
Agentic Coding
Aider Polyglot - coding benchmark evaluating model's ability to edit code across multiple programming languages.
SWE-bench Pro
Agentic Coding
SWE-bench Pro - refined benchmark with corrected problematic tasks in the public set. Internal agent scaffold; temp=1.0, top_p=0.95, 200K context window.
SWE Atlas - QnA
Agentic Coding
SWE Atlas QnA - agentic coding subtask; mini SWE agent harness, temp=1.0, top_p=0.95, 128K context, avg of 5 runs.
SWE-bench Verified
Agentic Coding
SWE-bench Verified evaluates LLM-based agents on real-world GitHub issue resolution from popular Python repositories. Internal agent scaffold (bash + file-ed...
Terminal Bench 2.1
Agentic Coding
Terminal Bench 2.1 - agentic coding and terminal use benchmark. Evaluated with Claude Code (avg@10), 5-hour timeout, max_tokens=131,072.
Terminal-Bench 2.1 (Terminus-2)
Agentic Coding
Terminal-Bench 2.1 evaluated with Harbor/Terminus-2 framework, parser=json, temp=1.0, top_p=1.0, 128K context, 4h timeout, 32 CPU/48GB RAM, avg of 5 runs.
SWE-Marathon
Agentic Coding
SWE-Marathon - long-horizon software engineering benchmark. Evaluated by Abundant AI with 1M context length, max effort level, and 128K maximum output tokens.
Terminal-Bench 2.1 (Claude Code)
Agentic Coding
Terminal-Bench 2.1 evaluated using Claude Code 2.1.126, parser=json, temp=1.0, top_p=1.0, max_new_tokens=131072, avg of 5 runs.
Terminal-Bench 3.0
Agentic Coding
Terminal-Bench 3.0 agentic terminal tasks, evaluated with the Claude Code 2.1.207 harness (reasoning effort=max, 400K context, 128K max output), avg@3 over t...