Claude Opus 4.8
Most capable Opus-tier model: state-of-the-art long-horizon agentic execution, knowledge work and memory. 1M context at standard pricing.
- name
- Claude Opus 4.8
- vendor
- Anthropic
- model_id
- claude-opus-4-8
- context_window
- 1M
- max_output
- 128K
- input_per_mtok
- $5.00
- output_per_mtok
- $25.00
- strengths
- Most capable Opus-tier model: state-of-the-art long-horizon agentic execution, knowledge work and memory. 1M context at standard pricing.
- provider
- Anthropic
- family
- Claude Opus
- release_date
- 2026-05-28 source
- last_updated
- 2026-06-15
- open_weights
- false source
- license
- proprietary source
- params_total
- — verify-against-primary-at-build ↗ Anthropic does not disclose parameter counts for Claude Opus 4.8.
- params_active
- — verify-against-primary-at-build ↗ Anthropic does not disclose parameter counts for Claude Opus 4.8.
- tool_call
- true source
- reasoning
- always source
- structured_output
- — verify-against-primary-at-build ↗ https://platform.claude.com/docs/en/about-claude/models/overview — schema-guaranteed structured-output flag not explicitly published for claude-opus-4-8 (no models.dev entry yet).
- attachment
- image source
- temperature
- — verify-against-primary-at-build ↗ https://platform.claude.com/docs/en/about-claude/models/overview — temperature-control flag not explicitly stated for claude-opus-4-8.
- knowledge_cutoff
- 2026-01 source
- context_advertised
- 1M source
- context_effective
- — verify-against-primary-at-build ↗ No published measured long-context recall benchmark for claude-opus-4-8; do not derive effective context from advertised.
- price_input
- $5.00 source
- price_output
- $25.00 source
- price_cache_read
- — verify-against-primary-at-build ↗ https://platform.claude.com/docs/en/about-claude/pricing — cache-read rate for claude-opus-4-8 not captured numerically in fetched sources.
- price_cache_write
- — verify-against-primary-at-build ↗ https://platform.claude.com/docs/en/about-claude/pricing — cache-write rate for claude-opus-4-8 not captured numerically in fetched sources.
- cost_per_full_window
- — verify-against-primary-at-build ↗ Derived at build: price_input × advertised context once cache context is confirmed.
- cost_per_agent_task
- — verify-against-primary-at-build ↗ Derived at build from the cache-aware agent-task model once cache rates are confirmed.
- modalities
- text-in,image-in,text-out source
- gpqa_diamond
- — verify-against-primary-at-build ↗ https://www.anthropic.com/news — benchmark number not captured from a primary leaderboard in this pass.
- swe_bench_verified
- — verify-against-primary-at-build ↗ https://www.swebench.com/ — SWE-bench Verified score for claude-opus-4-8 to confirm at build.
- terminal_bench
- — verify-against-primary-at-build ↗ https://www.tbench.ai/ — Terminal-Bench score for claude-opus-4-8 to confirm at build.
- tau2_bench
- — verify-against-primary-at-build ↗ Primary τ²-Bench leaderboard — score for claude-opus-4-8 to confirm at build.
- bfcl_tool_use
- — verify-against-primary-at-build ↗ https://gorilla.cs.berkeley.edu/leaderboard.html — BFCL tool-use score for claude-opus-4-8 to confirm at build.
- aa_index
- — verify-against-primary-at-build ↗ https://artificialanalysis.ai/ — Intelligence Index for claude-opus-4-8 to confirm at build.
- lmarena_elo
- — verify-against-primary-at-build ↗ https://lmarena.ai/leaderboard — human-preference Elo for claude-opus-4-8 to confirm at build.
- tokens_per_sec
- — verify-against-primary-at-build ↗ https://artificialanalysis.ai/ — measured throughput for claude-opus-4-8 to confirm at build.
- ttft
- — verify-against-primary-at-build ↗ https://artificialanalysis.ai/ — measured time-to-first-token for claude-opus-4-8 (overview lists Comparative latency: Moderate).
- hallucination_rate
- — verify-against-primary-at-build ↗ No primary hallucination benchmark captured for claude-opus-4-8 in this pass.
- agent_readiness_score
- — verify-against-primary-at-build ↗ Score withheld: R_tool (BFCL/τ²-Bench), R_ctx (effective context), R_cost (cache rates) and R_latency inputs not yet sourced. Compute per /models/agent-readiness-score once inputs confirmed.
- score_confidence
- partial
- source_url
- https://platform.claude.com/docs/en/about-claude/models/overview
- source_type
- provider_card
- last_verified
- 2026-06-15