Pass IndexThe State of AISign in

Claude 3.5 Sonnet

Claude 3.5 Sonnet is a model by Anthropic for writing code and reasoning and agentic, it stands 2nd of 217 on Mmlu — Epoch AI, and 4 sellers in the catalogue publish a price for it.

text + image + file → text + code · made by Anthropic

$3$15per Mtok in / out
AWS Bedrock · 4 sellers

Sold by 6 ways

SellerLaneRate
AWS Bedrock cloudstandard$3per Mtok in$15per Mtok out$0.3per Mtok cached$3.75per Mtok cache writeraw.githubusercontent.com · read 2026-09-16
Amazon Web Services cloudbatch$3per Mtok in$15per Mtok outaws.amazon.com · read 2026-08-25
DigitalOcean cloudstandard$3per Mtok in$15per Mtok out$0.3per Mtok cached$3.75per Mtok cache writemodels.dev · read 2026-09-16
Snowflake Cortex cloudstandard$3per Mtok in$15per Mtok out$0.3per Mtok cachedraw.githubusercontent.com · read 2026-09-16
AWS Bedrock cloudlong-context$6per Mtok in$30per Mtok out$0.6per Mtok cachedraw.githubusercontent.com · read 2026-09-16
Amazon Web Services cloudstandard$6per Mtok in$30per Mtok outaws.amazon.com · read 2026-08-25

Measured 29 standings

PlaceBoardMetricScore
2ndof 217Mmlu — Epoch AIEM0.873
7thof 15Cad eval — Epoch AIOverall pass (%)0.48
7thof 16The agent company — Epoch AI% Score0.344
12thof 22SWE-bench Multimodal% Resolved25.34
13thof 36Balrog — Epoch AIAverage progress0.326
14thof 22Cybench — Epoch AIUnguided % Solved0.175
20thof 69Aider polyglot coding leaderboardPercent correct64
20thof 54Live bench — Epoch AIGlobal average59.03
21stof 81Forecastbench — Epoch AIOverall score60.7
21stof 49Lech mazur writing — Epoch AIMean score8.03
23rdof 28LiveCodeBench Leaderboard (default window 8/1/2024–5/1/2025, 454 problems)Pass@136.4
29thof 38Gso — Epoch AIScore OPT@10.046
29thof 50Video mme — Epoch AIOverall (no subtitles)0.6
33rdof 38Vpct — Epoch AICorrect0.33
34thof 50Metr time horizons — Epoch AIaverage_score0.452
39thof 72Aider polyglot — Epoch AIPercent correct51.6
39thof 46Enigma eval — Epoch AIAccuracy0.91
49thof 51Humanity's Last Exam (Epoch AI replication)Accuracy4.08
55thof 108Math level 5 — Epoch AImean_score56.949
69thof 72Frontiermath tier 4 — Epoch AImean_score0
69thof 102Simplebench — Epoch AIScore (AVG@5)0.414
84thof 101FrontierMath (Tiers 1-3)mean_score2.069
92ndof 161Dtbench — Epoch AIAccuracy67.84
96thof 152LMArena · VisionScore (Elo)1161
110thof 170Weirdml — Epoch AIAccuracy39.97
180thof 402LMArena · TextScore (Elo)1374
218thof 313GPQA Diamondmean_score55.303
239thof 290OTIS Mock AIME 2024-2025mean_score8.472
508thof 656Epoch capabilities index — Epoch AIECI Score133.57

About

Claude 3.5 Sonnet — a text and code model from Anthropic, sold by 4 companies from $3 in and $15 out per million tokens, placed 2nd of 217 on Mmlu — Epoch AI.

It takes text, images and files and returns text and code, with a context window of 200,000 tokens. It was published in June 2024. Its sellers say it can call a tool. The catalogue files it under code, reasoning and agents. Four companies sell it. The cheapest is $3 in and $15 out per million tokens at AWS Bedrock. Beside the standard rate there are separately routed and batch lanes. It has been measured on 29 boards, and stands best at 2nd of 217 on Mmlu — Epoch AI.

Every current figure

Maker
Anthropic
Register
model
Takes
text + image + file
Returns
text + code
Context
200,000 tokens
Longest answer
8,200 tokens
Published
June 2024
Licence
not read
Sellers
4
Maker's own price
not read
Price
$3 in and $15 out per million tokens — AWS Bedrock
Boards
29
Best place
2nd of 217 — Mmlu — Epoch AI

Known as 3 names

DeepSeek R1 + claude-3-5-sonnet-20241022Claude-3.5-Sonnet-20241022Claude Sonnet 3.5