Claude Opus 5
- Lab
- anthropic
- Weights
- closed
- Released
- 2026-08-25
- Context window
- 1,000,000
- Modalities
- text, image
- API
- yes
- Local-runnable
- no
- Status
- current
announcementmodel cardpricing
Scores
Hard reasoning
| Value | Source | As of | Reported by | Notes |
| 63 index | artificial-analysis-intelligence-index ↗ | 2026-08-31 | aggregatorA | Adaptive Reasoning, Max Effort variant, ranked #1 of 178 models evaluated; Xhigh also 63, High 61. Index v4.1.1 (9 evals) — a different scale from the pre-v4.1.1 values stored on entity files. |
Human preference
| Value | Source | As of | Reported by | Notes |
| 1493 elo | lmarena-text ↗ | 2026-09-21 | aggregatorA | claude-opus-5-high, #10 on the text overall board; value unchanged from the 2026-09-14 fetch, as-of refreshed. Board states no date; today's fetch date used. |
| 1493 elo | lmarena-text ↗ | 2026-09-14 | aggregatorA | claude-opus-5-high variant, #10 on the text overall board; value unchanged from the 2026-09-07 fetch, as-of refreshed. |
| 1493 elo | lmarena-text ↗ | 2026-09-07 | aggregatorA | claude-opus-5-high, #9 on the text overall board; 1492 on the 2026-08-31 fetch. |
| 1492 elo | lmarena-text ↗ | 2026-08-31 | aggregatorA | claude-opus-5-high, #7 on the text overall board; 1493 on the 2026-08-26 fetch. |
| 1493 elo | lmarena-text ↗ | 2026-08-26 | aggregatorA | claude-opus-5-high variant on the 2026-08-26 board fetch; other axes pending weekly runs. |
Events
4 Anthropic released Claude Opus 5.5 (claude-opus-5-5) for long-running agentic coding and knowledge work, with a 1M-token default context window, 128k max output tokens, and always-on adaptive thinking (thinking cannot be disabled via the API; effort controls thinking depth instead). Pricing is $4/$20 per MTok versus Opus 5's $5/$25. Anthropic's own announcement states Opus 5.5 "performs at the level of Claude Fable 5.1 on most work and costs 40% less to run than Opus 5." Fast mode is available as a research preview on the Claude API, and computer-use on this model now requires a new toolset (computer_toolset_20260801) on the API and Google Cloud.
2026-09-22 Release Pricing change Tooling / agents Claude Platform release notes 4 On September 1, 2026, Anthropic released Claude Fable 5.1 and Claude Mythos 5.1, positioned for long-running agentic coding, knowledge work, and research, both with a 1M-token context window, 128k max output tokens, and always-on adaptive thinking. List pricing stays at $10/$50 per MTok input/output (same as Fable 5), but prompt cache read price drops 75% to $0.25/MTok. Anthropic also introduced Enterprise Frontier Safeguards (EFS) for agent observability. Per Stratechery, Fable's prior data retention policy was removed rather than merely altered. Anthropic's own benchmark table reports Fable 5.1 scoring 52.6% on the new Terminal-Bench-Science 0.1 benchmark versus 24.7% for Fable 5, 29.0% for Opus 5, and 22.4% for GPT-5.6 Sol. TechCrunch reported the release also reduces false-positive safeguard restrictions. Per Latent Space's AINews recap, Artificial Analysis measured roughly 1.7x more output tokens per task, offsetting the cache savings for a reported ~20% net per-task cost increase.
← Back to Models