Registry models

Registry models
ModelProviderReleasedStatusBenchmark score
Claude Haiku 5.5AnthropicOctober 7, 2026activeCursorBench 4.0 (high)42.3%
Gemini 4 ArgonGoogle (DeepMind)September 30, 2026previewAPEX-Agents Original (high)82.4%
GPT-6.1 SolOpenAISeptember 29, 2026activeAPEX-Agents Original (max)60.0%
Claude Sonnet 5.5AnthropicSeptember 28, 2026activeAPEX-Agents Original (max)75.5%
Claude Opus 5.5AnthropicSeptember 22, 2026activeAPEX-Agents Original (max)73.5%
GPT-6 LunaOpenAISeptember 22, 2026activeAPEX-Agents Original (max)44.3%
GPT-6 SolOpenAISeptember 22, 2026activeAPEX-Agents Original (max)54.3%
Grok 4.7SpaceXAISeptember 21, 2026activeAPEX-Agents Original (xhigh)54.6%
GPT-6 AstraOpenAISeptember 3, 2026activeAPEX-Agents Original (max)64.7%
Gemini 3.8 FlashGoogle (DeepMind)September 2, 2026activeAPEX-Agents Original (high)64.3%
Muse Spark 1.3Meta AI (originally Facebook AI Research)September 2, 2026activeAPEX-Agents Original (max)47.6%
Claude Fable 5.1AnthropicSeptember 1, 2026activeAPEX-Agents Original (high)59.7%
GLM-5.3-FlashZ.aiAugust 26, 2026activeAPEX-Agents Original (max)52.8%
GLM-5.3Z.aiAugust 14, 2026activeAPEX-Agents Original (max)56.6%
Gemini 3.7 FlashGoogle (DeepMind)August 13, 2026activeAPEX-Agents Original (high)67.8%
Grok 4.6SpaceXAIAugust 12, 2026activeAPEX-Agents Original (xhigh)65.3%
MAI-Thinking-1Microsoft AIAugust 12, 2026activeRiemann-bench October 2026 leaderboard4.8%
Muse Glimmer 30BMeta AI (originally Facebook AI Research)August 10, 2026activeChartography October 2026 leaderboard (xHigh)17.2%
Muse Spark 1.2Meta AI (originally Facebook AI Research)August 5, 2026activeAPEX-Agents Original (xhigh)36.4%
Qwen3.8-MaxAlibaba (Tongyi)August 2, 2026activeAPEX-Agents Original (xhigh)63.3%
Inkling-SmallThinking Machines LabJuly 30, 2026activeChartography October 2026 leaderboard (xHigh)9.1%
Claude Opus 5AnthropicJuly 24, 2026activeAPEX-Agents Original (max)65.8%
Gemini 3.6 FlashGoogle (DeepMind)July 21, 2026activeAPEX-Agents Original (high)46.9%
Grok 4.5SpaceXAIJuly 16, 2026activeAPEX-Agents Original (high)56.2%
Kimi K3Moonshot AIJuly 16, 2026activeAPEX-Agents Original (max)50.6%
InklingThinking Machines LabJuly 15, 2026activeAPEX-Agents Original (high)33.8%
GPT-5.6 LunaOpenAIJuly 9, 2026activeAPEX-Agents Original (max)43.0%
GPT-5.6 SolOpenAIJuly 9, 2026activeAutomationBench 1.0.63.8%
GPT-5.6 TerraOpenAIJuly 9, 2026activeAPEX-Agents Original (max)58.2%
Muse Spark 1.1Meta AI (originally Facebook AI Research)July 9, 2026activeAPEX-Agents Original (xhigh)31.8%
Claude Sonnet 5AnthropicJune 30, 2026activeAPEX-Agents Original (max)54.5%
GLM-5.2Z.aiJune 16, 2026activeAPEX-Agents Original (max)45.2%
Claude Fable 5AnthropicJune 9, 2026activeAPEX-Agents Original (max)63.6%
NVIDIA Nemotron 3 Ultra 550B-A55BNVIDIAJune 4, 2026activeAPEX-Agents Original (high)22.7%
Gemma 4 12B InstructGoogle (DeepMind)June 3, 2026activeBIG-Bench Extra Hard Original — micro average (thinking mode)53.0%
Qwen3.7-PlusAlibaba (Tongyi)June 3, 2026activeAutomationBench 1.0.63.7%
MAI-Code-1-FlashMicrosoft AIJune 2, 2026activeGPQA Diamond84.6%
MiniMax M3MiniMaxJune 1, 2026activeAPEX-Agents Original (high)37.7%
Claude Opus 4.8AnthropicMay 28, 2026activeAPEX-Agents Original (max)48.9%
Mistral Medium 3.5Mistral AIMay 22, 2026activeFinance Agent Benchmark 2 — October 2026 evaluation (high)32.1%
Qwen3.7-MaxAlibaba (Tongyi)May 20, 2026activeGPQA Diamond — Epoch AI harness90.9%
Gemini 3.5 FlashGoogle (DeepMind)May 19, 2026activeAPEX-Agents Original (high)27.5%
Composer 2.5CursorMay 18, 2026activeVibe Code Bench 1.1 — Cursor CLI, October 2026 evaluation49.6%
DeepSeek-V4-FlashDeepSeek AIApril 24, 2026activeAutomationBench 1.0.6 (max)18.1%
DeepSeek-V4-ProDeepSeek AIApril 24, 2026activeBrowseComp 130-question subset — Mercor web research agent (max)62.3%
GPT-5.5OpenAIApril 23, 2026activeAPEX-Agents Original (xhigh)55.1%
Kimi K2.6Moonshot AIApril 20, 2026activeAutomationBench 1.0.62.6%
Qwen3.6-35B-A3BAlibaba (Tongyi)April 17, 2026activeGPQA Diamond — Epoch AI harness83.8%
Claude Opus 4.7AnthropicApril 16, 2026activeAPEX-Agents Original (max)49.2%
Muse SparkMeta AI (originally Facebook AI Research)April 8, 2026activeGPQA Diamond — Epoch AI harness89.8%