anthropic/claude-opus-5| Provider | Input, $ per 1M | Output, $ per 1M | In our data since |
|---|---|---|---|
| google-vertex-anthropic | $5 | $25 | 31 Jul 2026 |
Amazon Bedrock
anthropic.claude-opus-5
|
$5 | $25 | 31 Jul 2026 |
| Anthropic | $5 | $25 | 31 Jul 2026 |
| Azure Cognitive Services | $5 | $25 | 31 Jul 2026 |
| Cloudflare AI Gateway | $5 | $25 | 31 Jul 2026 |
| CrossModel | $5 | $25 | 31 Jul 2026 |
| GitHub Copilot | $5 | $25 | 31 Jul 2026 |
| Google Vertex AI | $5 | $25 | 31 Jul 2026 |
| LLM Gateway | $5 | $25 | 31 Jul 2026 |
| Microsoft Foundry | $5 | $25 | 31 Jul 2026 |
| OpenCode Zen | $5 | $25 | 31 Jul 2026 |
| OpenRouter | $5 | $25 | 31 Jul 2026 |
| Vercel | $5 | $25 | 31 Jul 2026 |
Amazon Bedrock
eu.anthropic.claude-opus-5
|
$5.5 | $27.5 | 31 Jul 2026 |
| Venice AI | $6 | $30 | 31 Jul 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.
| Task set | Result | Run conditions | Measured by |
|---|---|---|---|
| Arena Score, web development | 1,704.67 1,689–1,720 | effort: max | — |
| Arena Score, mathematics | 1,544.36 1,507–1,582 | effort: high | — |
| Arena Score, programming | 1,534.15 1,510–1,559 | effort: max | — |
| Arena Score, hard prompts | 1,532.57 1,517–1,548 | effort: max | — |
| Arena Score, expert questions | 1,524.59 1,486–1,563 | effort: max | — |
| Arena Score in French | 1,522.24 1,479–1,565 | effort: high | — |
| Arena Score, instruction following | 1,519.69 1,498–1,541 | effort: max | — |
| Arena Score, overall | 1,511.6 1,499–1,524 | effort: max | — |
| Arena Score, creative writing | 1,510.88 1,481–1,541 | effort: max | — |
| Arena Score, long queries | 1,506.13 1,488–1,525 | effort: max | — |
| Arena Score in Russian | 1,503.36 1,463–1,544 | effort: max | — |
| Arena Score in English | 1,500.16 1,482–1,519 | effort: max | — |
| Arena Score, multi-turn dialogue | 1,493.98 1,460–1,528 | effort: max | — |
| Arena Score, text recognition in images | 1,332.16 1,315–1,349 | effort: high | — |
| Arena Score, working with images | 1,328.68 1,314–1,343 | effort: high | — |
| Arena Score, understanding diagrams | 1,323.16 1,295–1,352 | effort: high | — |
| Mock AIME 2024–2025 — olympiad problems | 98.89 % | effort: max · with a tuned harness | Epoch evaluations |
| ARC-AGI — generalising to unseen patterns | 97.5 % | effort: max · with a tuned harness | https://arcprize.org/leaderboard |
| GPQA Diamond — graduate-level questions | 91.84 % | effort: max · with a tuned harness | Epoch evaluations |
| WeirdML — unusual machine learning tasks | 91.78 % | effort: max | https://htihle.github.io/weirdml.html |
| ARC-AGI-2 | 88.33 % | effort: max | — |
| FrontierMath, levels 1–3 | 85.61 % | effort: max | Epoch evaluations |
| FrontierMath, level 4 — research-grade problems | 73.17 % | effort: max | Epoch evaluations |
| CursorBench — edits in the editor | 70 % | effort: max | https://cursor.com/cursorbench |
| SimpleQA Verified — factual accuracy | 56.7 % | effort: max | Epoch evaluations |
| FrontierCode — patches fit to be merged into a project | 53.4 % | effort: max | https://cognition.com/frontiercode |
| APEX-Agents | 43.5 % | effort: max | — |
| Chess puzzles | 38.97 % | effort: max | Epoch evaluations |
| CritPt — physics problems | 29.14 % | effort: max | — |
| Arena Score, agent tasks | 0.12 0–0 | effort: max | — |