Best LLMs for long-horizon engineering and research

FrontierSWE

Tests substantial engineering and research assignments, helping you assess models for complex client work that extends beyond a quick code change. Each of the 34 tasks has a 20-hour trial budget; scores do not translate directly into billable hours saved.

Updated: Nov 9, 2026Data as of: Nov 9, 2026

Limited reportingNot saturated

Official benchmark

FrontierSWE results. Each row identifies its benchmark version and evaluation track. Scores from different versions are not directly comparable. Best effort from the latest stored report date per model and version.
Model / creator	Version	Reported score
Claude Fable 5.2Anthropic · Proprietary	2 / Proximal harness	69.9%
GPT-6 AstraOpenAI · Proprietary	2 / Proximal harness	65.5%
Claude Opus 5.5Anthropic · Proprietary	2 / Proximal harness	62.3%
Gemini 4 CarbonGoogle · Proprietary	2 / Proximal harness	61.2%
Claude Fable 5.1Anthropic · Proprietary	2 / Proximal harness	56.3%

View all 5 published reports

Claude Fable 5.269.9%

2 / Proximal harness

effort: max · Proximal harness; 34 tasks; 20h/trial; mean of five trials. · Tools enabled

Anthropic ↗

GPT-6 Astra65.5%

2 / Proximal harness

effort: highest available · November 9 launch-table comparison; highest available effort unless explicitly noted. Preserve the benchmark track and model configuration; competitor figures are as reported by Anthropic. · Tools enabled

Anthropic ↗

Claude Opus 5.562.3%

2 / Proximal harness

Anthropic ↗

Gemini 4 Carbon61.2%

2 / Proximal harness

Anthropic ↗

Claude Fable 5.156.3%

2 / Proximal harness

Anthropic ↗

Sources and original reports ↗

Other benchmarks

BrowseComp

Tests whether an agent can find hard-to-locate, verifiable information online, helping you choose a model for client research that would otherwise require persistent manual searching.

SWE-bench Pro

Tests substantial software repository issues, helping you assess which model could support complex fixes and implementation work for clients.

DeepFinance

Tests private equity and investment banking work such as valuation, due diligence and financial modeling, helping you assess AI support for demanding client analysis.