Model releases1 mins read
Introducing new APEX-Accounting benchmark, built with Ramp
APEX-Accounting is live. No frontier model can reliably close the books.
Model releases1 mins read
Opus 5 tops the APEX-Agents leaderboard
Claude Opus 5 debuts at #1 on APEX-Agents and takes the top spot on SWE Integration.
Model releases3 mins read
Grok 4.5 and GPT-5.6 land on APEX
Grok 4.5 debuts at #2 on APEX-SWE and takes the overall Integration lead.
Model releases1 mins read
Fable 5 is back. Here's what changed.
Fable 5's re-release scores 54.8% on APEX-SWE, still 9.5 points ahead of Opus 4.8.
Model releases1 mins read
Claude Sonnet 5 earns #3 spot on APEX-SWE and makes APEX-Agents top 10
Claude Sonnet 5 ranks #3 on APEX-SWE (43.7%) and jumps into the APEX-Agents top 10.
Model releases1 mins read
Claude Fable 5 tops APEX-SWE with a 20-point lead
Anthropic's Claude Fable 5 was released this week. We tested it on APEX-Agents and APEX-SWE ahead of the launch.
Model releases1 mins read
Claude Opus 4.8 tops APEX-SWE, places 2nd on APEX-Agents
Anthropic's Claude Opus 4.8 was released yesterday. We tested it on APEX-Agents and APEX-SWE ahead of the launch.
Model releases1 mins read
Gemini 3.5 Flash tops all APEX-Agents leaderboards
Google's Gemini 3.5 Flash (High) took the top spot on the APEX-Agents leaderboard with 49.6% at Pass@1.