all the models — AI benchmark observatory
← Developer AI tools

Agent · CLI

Devin

Autonomous coding agent with Adaptive routing and a multi-lab model menu.

Desktop/CLI pickers include Claude, GPT, Gemini, Grok, Kimi, GLM, and DeepSeek plus Cognition’s SWE family.

OpenAI13Anthropic10Google6Cognition4Devin4xAI4Moonshot3DeepSeek2Z.ai2NVIDIA1Thinking Machines1

Available models

50 listed · sourced 2026-09-08 from vendor docs

ModelProviderNotes
AdaptiveDevinRecommended router — picks model per task
Fast ArenaDevinArena mode battle group
Hybrid ArenaDevinArena mode battle group
Frontier ArenaDevinArena mode battle group
SWE-1.7CognitionMax and Medium effort variants
SWE-1.7 LightningCognitionSWE-1.7 on Cerebras
SWE-1.6Cognition
SWE-1.6 FastCognition
Claude Opus 5AnthropicEffort / Fast variants in picker
Claude Opus 4.8AnthropicEffort / Fast variants in picker
Claude Opus 4.7Anthropic
Claude Opus 4.6Anthropic
Claude Fable 5.1Anthropic
Claude Fable 5Anthropic
Claude Sonnet 5Anthropic
Claude Sonnet 4.6Anthropic
Claude Sonnet 4.5Anthropic
Claude Haiku 4.5Anthropic
GPT-5.6 SolOpenAIDevin Local for GPT-5.6 variants
GPT-5.6 TerraOpenAIDevin Local
GPT-5.6 LunaOpenAIDevin Local
GPT-5.5OpenAIEffort / Fast variants in picker
GPT-5.4OpenAIIncludes Mini; effort / Fast variants
GPT-5.3 CodexOpenAIIncludes Codex Spark
GPT-5.2OpenAIEffort / Fast variants in picker
GPT-5.1OpenAI
GPT-5OpenAI
GPT-4.1OpenAI
GPT-4oOpenAI
o3OpenAI
GPT-OSS 120BOpenAI
Gemini 3.7 FlashGoogleIntroductory pricing through Dec 31, 2026
Gemini 3.6 FlashGoogle
Gemini 3.5 FlashGoogle
Gemini 3.1 ProGoogle
Gemini 3 FlashGoogle
Gemini 2.5 ProGoogle
Grok 4.6xAI
Grok 4.5xAI
Grok Code Fast 1xAI
Grok 3xAI
Kimi K3Moonshot
Kimi K2.7Moonshot
Kimi K2.6Moonshot
GLM-5.3Z.aiIncludes Flash and 1M variants
GLM-5.2Z.aiIncludes 1M variants
DeepSeek V4 ProDeepSeek
DeepSeek V4 FlashDeepSeek
InklingThinking MachinesEffort variants in picker
Nemotron 3 UltraNVIDIA

Other multi-vendor tools