What the vendor claims, next to what independent boards measured. The two are never mixed in one column.
Source: OpenRouter, as of Oct 5, 03:00 UTC. Model page on OpenRouter
Quoted word for word from the vendor's own launch post or model card. Claims, not measurements.
No vendor claims recorded yet. They come from launch posts and model cards the Drops covered.
Measured by Epoch AI or LMArena, not by the vendor. "Listed as" is the source's own name for the model, so you can check our match.
| Board | Result | Rank |
|---|---|---|
| Epoch AI · FrontierMath-Tiers-1-3-v2-PrivateListed as: claude-opus-5-5_maxSource: Epoch AI (CC BY), as of Oct 5, 2026. Source | 91.2%Accuracy | 3 |
| Epoch AI · FrontierMath-Tier-4-v2-PrivateListed as: claude-opus-5-5_maxSource: Epoch AI (CC BY), as of Oct 5, 2026. Source | 95.0%Accuracy | 6 |
| Epoch AI · SimpleQA VerifiedListed as: claude-opus-5-5_maxSource: Epoch AI (CC BY), as of Oct 5, 2026. Source | 72.2%Accuracy | 4 |
| Epoch AI · OTIS Mock AIME 2024-2025Listed as: claude-opus-5-5_maxSource: Epoch AI (CC BY), as of Oct 5, 2026. Source | 100.0%Accuracy | 1 |
| Epoch AI · Epoch Capabilities IndexListed as: Claude Opus 5.5Source: Epoch AI (CC BY), as of Oct 5, 2026. Source | 167.35Index | 1 |
| LMArena · Text (style control)Listed as: claude-opus-5.5-highLMArena (CC BY 4.0), as of Oct 2, 2026. Source | 1,504Arena score | 4 |
| LMArena · AgentListed as: Claude Opus 5.5 (High)LMArena (CC BY 4.0), as of Oct 2, 2026. Source | 0.138Score | 2 |
On LMArena's WebDev board as of 2026-10-01, claude-opus-5.5-max scores 1815.361, ahead of gpt-6-astra-max at 1787.666 and claude-sonnet-5.5-xhigh at 1786.263. Source: LMArena (CC BY 4.0).
Why it matters: WebDev is the board closest to front-end work, and Opus 5.5 leads GPT-6 Astra there; worth a test if web apps are your main use.
From Drop #005 · Mon, Oct 5 · MorningEpoch AI's Capabilities Index ranks Claude Opus 5.5 first at 167.35, GPT-6 Astra second at 166.51 and Claude Sonnet 5.5 third at 165.2. Anthropic models hold four of the top five spots.
Why it matters: The top three are close, so choosing a frontier model comes down to price, speed and fit for your own tasks.
From Drop #004 · Sun, Oct 4 · EveningThe guide says Opus 5.5 thinks before every reply and runs longer on its own, so state what done looks like, set stop rules in CLAUDE.md and keep permission prompts on for destructive commands.
Why it matters: The guide says Opus 5.5 is the first Opus with Fable-level bio and cyber safeguards; flagged messages can move to an older model. The HN thread drew 124 comments.
From Drop #003 · Sun, Oct 4 · Morning