Anthropic

Claude Opus 4.6

Type
language model
Context
1,000K tokens
Max output
128K
Released
5 February 2026
Knowledge cutoff
May 2025
API string
anthropic/claude-opus-4-6

Prices by provider

Provider Input, $ per 1M Output, $ per 1M In our data since
Kilo Code stealth/claude-opus-4.6 $4 $20 31 Jul 2026
Poe $4.3 $21 31 Jul 2026
NanoGPT $4.998 $25.007 31 Jul 2026
neon $5 $25 31 Jul 2026
frogbot $5 $25 31 Jul 2026
auriko $5 $25 31 Jul 2026
google-vertex-anthropic $5 $25 31 Jul 2026
freemodel $5 $25 31 Jul 2026
perplexity-agent $5 $25 31 Jul 2026
vercel_ai_gateway $5 $25 31 Jul 2026
302.AI $5 $25 31 Jul 2026
AIHubMix $5 $25 31 Jul 2026
Abacus.AI $5 $25 31 Jul 2026
Amazon Bedrock anthropic.claude-opus-4-6-v1 $5 $25 2 Aug 2026
Anthropic $5 $25 31 Jul 2026
Azure Cognitive Services $5 $25 31 Jul 2026
Cloudflare AI Gateway $5 $25 31 Jul 2026
Databricks $5 $25 31 Jul 2026
DigitalOcean Gradient AI $5 $25 1 Aug 2026
GMI Cloud $5 $25 31 Jul 2026
GitHub Copilot $5 $25 31 Jul 2026
Google Vertex AI $5 $25 31 Jul 2026
Jiekou.AI $5 $25 31 Jul 2026
Kilo Code anthropic/claude-opus-4.6 $5 $25 31 Jul 2026
LLM Gateway $5 $25 31 Jul 2026
Merge Gateway $5 $25 31 Jul 2026
Microsoft Foundry $5 $25 31 Jul 2026
NEAR AI $5 $25 31 Jul 2026
OpenCode Zen $5 $25 31 Jul 2026
OpenRouter $5 $25 31 Jul 2026
OrcaRouter $5 $25 31 Jul 2026
Pioneer $5 $25 31 Jul 2026
Requesty $5 $25 31 Jul 2026
SAP AI Core $5 $25 1 Aug 2026
Vercel $5 $25 31 Jul 2026
ZenMux $5 $25 31 Jul 2026
Amazon Bedrock eu.anthropic.claude-opus-4-6-v1 $5.5 $27.5 2 Aug 2026
Venice AI $6 $30 31 Jul 2026
Amazon Bedrock au.anthropic.claude-opus-4-6-v1 $16.5 $82.5 2 Aug 2026

Only the base tier and only per-token prices are shown. Batch, discounted and cached rates, as well as prices per image or per second of video, do not go into this table: they cannot stand in the same column as a price per million tokens.

A single provider appears several times if it sells this model under different identifiers and at different prices — for example, at different weight precision. The identifier is shown next to the name. Identical offers that arrived from two sources under different spellings are merged into one row.

The date in the last column is the day this price first entered our collection. The price may well be older: before that day we simply were not recording it. It has not changed since — otherwise a new row with a new date would stand in its place.

Measurement results

Task set Result Run conditions Measured by
Arena Score, web development 1,537.93 1,532–1,544
Arena Score, expert questions 1,536.69 1,528–1,545
Arena Score, programming 1,534.96 1,529–1,541
Arena Score, hard prompts 1,521.95 1,518–1,526
Arena Score, long queries 1,515.24 1,510–1,520
Arena Score in Spanish 1,511.71 1,497–1,526
Arena Score, instruction following 1,508.91 1,504–1,514
Arena Score, mathematics 1,507.89 1,498–1,518
Arena Score, multi-turn dialogue 1,506.79 1,500–1,513
Arena Score in English 1,505.8 1,501–1,510
Arena Score in Russian 1,498.93 1,491–1,507
Arena Score in French 1,497.74 1,483–1,513
Arena Score, overall 1,496.63 1,493–1,500
Arena Score, creative writing 1,481.69 1,475–1,489
Arena Score, understanding diagrams 1,339.65 1,330–1,349
Arena Score, text recognition in images 1,325.11 1,319–1,332
Arena Score, working with images 1,312.19 1,306–1,319
Mock AIME 2024–2025 — olympiad problems 94.44 % thinking budget: 64K · with a tuned harness Epoch evaluations
ARC-AGI — generalising to unseen patterns 94 % thinking budget: 120K · with a tuned harness ARC Prize Leaderboard
Cybench — cybersecurity tasks 93 % · with a tuned harness Opus 4.6 System Card self-reported
GPQA Diamond — graduate-level questions 87.37 % thinking budget: 32K · with a tuned harness Epoch evaluations
Terminal-Bench — working in the command line 79.8 % · with a tuned harness https://www.tbench.ai/leaderboard/terminal-bench/2.0
SWE-bench Verified — fixing bugs in repositories 78.72 % · with a tuned harness Epoch evaluations
WeirdML — unusual machine learning tasks 77.95 % effort: high https://htihle.github.io/weirdml.html
ARC-AGI-2 69.17 % thinking budget: 120K
FrontierMath, levels 1–3 65.97 % effort: max Epoch evaluations
SimpleBench — trick questions 61.12 % SimpleBench Leaderboard
DeepResearch Bench — deep research 55.31 % effort: high https://drb.futuresearch.ai/#drb self-reported
SimpleQA Verified — factual accuracy 46.49 % effort: max Epoch evaluations
GSO-Bench — code optimisation 41.2 % effort: high GSO Leaderboard
APEX-Agents 32.4 %
Humanity’s Last Exam — expert-level questions 31.13 % effort: max
FrontierMath, level 4 — research-grade problems 26.83 % effort: max Epoch evaluations
Chess puzzles 12.67 % thinking budget: 120K Epoch evaluations
Remote Labor Index — jobs from a freelance marketplace 4.17 %
Arena Score, agent tasks 0.07 0–0