Anthropic · proprietary

Claude Opus 4.8

Claude Opus 4.8 by Anthropic appears in 1 source with Reasoning at 65.69. Best read for LLM, Long context, Low cost.

CreatorAnthropic

Release date2026-05-28

Knowledge cutoffNot published

Context1M tokens

Input price$5/M tokens

Output price$25/M tokens

Modalitytext + vision

CountryUS

Metrics

All source-backed metrics

Open in leaderboard

LLM Stats Rank63ranking · source_rank

Reasoning65.69reasoning · index_reasoning

Math54.54math · index_math

Coding52.28coding · index_code

Research39.16research · index_search

Vision48.23multimodal · index_vision

Tool calling41.61tool_calling · index_tool_calling

Long context34.13long_context · index_long_context

Finance31.22domain · index_finance

Legal41.19domain · index_legal

GPQA93.6 %reasoning · gpqa_score

SWE-bench Verified88.6 %coding · swe_bench_verified_score

Code Arena805.44 %coding · coding_arena_score

Humanity Last Exam57.9 %reasoning · hle_score

BrowseComp84.3 %research · browsecomp_score

Toolathlon59.9 %tool_calling · toolathlon_score

MCP Atlas82.2 %tool_calling · mcp_atlas_score

SWE-bench Pro69.2 %coding · swe_bench_pro_score

CharXiv-R89.9 %multimodal · charxiv_r_score

ScreenSpot Pro87.9 %multimodal · screenspot_pro_score

Context1,000,000 tokenscontext · context

Speed116.74 c/sperformance · throughput

Latency1,665.12 msperformance · latency

Input price5 $/Mpricing · input_price

Output price25 $/Mpricing · output_price

Evidence

Citations and source overlap

LLM StatsPublic full LLM leaderboard advanced explorer snapshot, including ranking, benchmark indexes, pricing, speed, context, license, and model metadata.Retrieved 2026-06-01

FAQ

How should I read this profile?

Treat this as a source-backed model dossier, not an EvalKit-run verification. The public values are replicated from linked sources and kept source-scoped.

Is Claude Opus 4.8 verified by EvalKit?

No. EvalKit currently shows 0 verified rows until real run evidence exists.

Why can metrics disagree?

Different sources test different tasks, dates, prompts, and aggregation methods. EvalKit keeps those differences visible instead of merging them into a fake universal score.