AnthropicReleased Mar 4, 2024
Proprietary
opus-3
Its best results are in Abstract Reasoning (#182 of 201) and Reliability (#173 of 177); weakest in Mathematical Reasoning (#175 of 176). Three of seven dimensions are estimated, because no results are published yet.
Composite IQ ?
80
Effective cost
$51.65/1M
$15 in · $75 out
Speed
Not measured
No speed data yet
Context
Unverified
Not in publisher docs
Evidence
4of 7 dims
7 of 39 benchmarks
IQ profile
Seven dimensions, averaged equally into Composite IQ
MeasuredEstimated95% range
#1 in dimensionTop-10 median
IQ dimensions
Ranks are among models scored in each dimension. Ranges are model-based estimates, not calibrated confidence intervals.
DimensionIQPosition · 70–160RankEvidencevs Top 10
opus-395% range (model-based, uncalibrated)Top-10 median#1 in dimension100 = average human
Benchmark results
7 of 39 Composite IQ benchmarks have a published result for opus-3. Benchmarks without a result stay listed so gaps are visible.
Abstract Reasoning IQ 75 · measuredBest: #74 of 76 on Chess Puzzles1/5
BenchmarkScorePercentileRankLeader · source
Chess Puzzles
Epoch AI vEpoch 100-position leaderboard (task implementation versions recorded per run) · Oct 14
5.0%
#74 / 76
Leader gpt-6-astra · 72.0%
Epoch AI vEpoch 100-position leaderboard (task implementation versions recorded per run) · Oct 14
Epoch AI vEpoch 100-position leaderboard (task implementation versions recorded per run) · Oct 14
Not yet reported
ARC-AGI-3ARC-AGI-2ARC-AGI-1Mystery Game Puzzles
Mathematical Reasoning IQ 71 · measuredBest: #56 of 56 on AIME1/5
BenchmarkScorePercentileRankLeader · source
Not yet reported
FrontierMath Tier 4FrontierMath Tier 1-3ProofBench 1.1MathArena
Academic Reasoning IQ 92 · measuredBest: #43 of 45 on MMLU-Pro4/6
BenchmarkScorePercentileRankLeader · source
MMLU-Pro
TIGER-Lab MMLU-Pro leaderboard (missing rows only) · Sep 22
68.5%
#43 / 45
Leader fable-5.1 · 92.4%
TIGER-Lab MMLU-Pro leaderboard (missing rows only) · Sep 22
TIGER-Lab MMLU-Pro leaderboard (missing rows only) · Sep 22
SciCode
Artificial Analysis model leaderboard · Sep 10
23.3%
#163 / 167
Leader opus-5.5 · 66.9%
Artificial Analysis model leaderboard · Sep 10
Artificial Analysis model leaderboard · Sep 10
GPQA Diamond
Artificial Analysis v4.3.2 · Oct 27
48.9%
#144 / 147
Leader gpt-6-astra · 96.1%
Artificial Analysis v4.3.2 · Oct 27
Artificial Analysis v4.3.2 · Oct 27
Humanity's Last Exam
Artificial Analysis v4.3.2 · Oct 27
2.8%
#166 / 167
Leader opus-5.5 · 61.4%
Artificial Analysis v4.3.2 · Oct 27
Artificial Analysis v4.3.2 · Oct 27
Not yet reported
CritPtMMMU-Pro
Programmatic Reasoning IQ 79 · estimatedNo published results0/6
No published results yet for Programmatic Reasoning. The 79 shown is estimated from opus-3's other dimensions and will be replaced as results come in. How estimates work
LiveCodeBenchIOI V2Terminal-Bench 4.0ProgramBenchSWE-rebenchMirrorCode
Computer Use IQ 75 · estimatedNo published results0/5
No published results yet for Computer Use. The 75 shown is estimated from opus-3's other dimensions and will be replaced as results come in. How estimates work
BrowseCompOSWorld-VerifiedMCP AtlasArena.ai Agent ArenaAgents' Last Exam
Reliability IQ 86 · measuredBest: #77 of 80 on SimpleQA Verified1/6
BenchmarkScorePercentileRankLeader · source
SimpleQA Verified
Full source leaderboard reconciliation · Oct 23
12.6%
#77 / 80
Leader Gemini 4 Argon · 76.0%
Full source leaderboard reconciliation · Oct 23
Full source leaderboard reconciliation · Oct 23
Not yet reported
aaOmniscienceBullshitBench v2IFBenchAA Long Context Reasoning v1.1FACTS Grounding
Knowledge Work IQ 79 · estimatedNo published results0/6
No published results yet for Knowledge Work. The 79 shown is estimated from opus-3's other dimensions and will be replaced as results come in. How estimates work
AA-Briefcase v1.1GDPval-AA v2.1AutomationBench-AAGDP.pdfAA-AnalystAgentEnterpriseOps-Gym-AA
Release & availability
Release details on record for this model.
- Released
- Developer
- Anthropic
- Weights
- ProprietaryHosted access only
- API price
- $15 in · $75 outper 1M tokens · full pricing
About this score
How much of opus-3's IQ is measured, and how the rest is estimated.
7/ 39benchmark results
4/ 7dimensions measured
3estimated
1/5Abstract
1/5Math
4/6Academic
0/6Coding
0/5Computer Use
1/6Reliability
0/6Knowledge Work
Bar height = share of benchmarks with resultsEstimatedNo score
Measured results are shown as published
Every score in the benchmark tables comes from its source unchanged. Composite IQ averages all seven dimensions equally.
Estimates fill the gaps
Programmatic Reasoning, Computer Use and Knowledge Work have no published results yet, so their IQ is estimated from opus-3's other dimensions.
A missing result is not a zero
Benchmark versions and evaluation harnesses can differ between sources, so small gaps between models may not be meaningful.
Compare opus-3
Side-by-side pages for the models it is most often weighed against.
#1 overall
gpt-6.1-sol
IQ141 61 higherEffective cost$6.30 8.2× cheaper
Anthropic #1
opus-5.5
IQ141 61 higherEffective cost$35.73 1.4× cheaper
Closest IQ
GPT-4.5 Preview
IQ82 2 higher
Most similar profile
gpt-4-turbo
IQ78 2 lowerEffective cost$50.29 about the same
Similar price
Fugu Ultra v2.0
IQ109 29 higherEffective cost$44.78 1.2× cheaper
#3 overall
Gemini 4 Argon
IQ140 60 higher
opus-3Comparison model
Frequently asked questions
Quick answers about opus-3, drawn from the same data as the rest of this page.
How AI IQ scores models →Get this data from the API →- What is the IQ of opus-3?
- opus-3 has an estimated IQ of 80, which ranks #175 out of 176 public models on the AI IQ leaderboard. The score combines seven equally weighted reasoning dimensions using measured benchmark results and conservative estimates for missing coverage.
- What is opus-3 best at?
- Among the dimensions with published benchmark results, opus-3 ranks highest in Abstract Reasoning (IQ 75, #182 of 201) and Reliability (IQ 86, #173 of 177). Its lowest-ranked measured dimension is Mathematical Reasoning (IQ 71, #175 of 176).
- What is the EQ of opus-3?
- opus-3 doesn't have an EQ score yet. EQ (emotional reasoning) is calculated from EQ-Bench 4 and AttuneBench, and neither has published a result for opus-3. AI IQ doesn't estimate EQ from other benchmarks, so the score appears once one of those results is published.
- Which benchmarks has opus-3 been tested on?
- opus-3 has published results on 7 of the 39 benchmarks behind AI IQ's Composite IQ, including MMLU-Pro, SimpleQA Verified, Chess Puzzles, SciCode and GPQA Diamond.
- How much does opus-3 cost?
- opus-3 costs $15.00 per million input tokens and $75.00 per million output tokens. Adjusted for how many tokens it actually consumes on real tasks, its effective cost works out to $51.65 per million input-plus-output tokens.
- Is opus-3 open source?
- No — opus-3 is a proprietary model. Its weights are not publicly available, so it is used through hosted APIs and products.
- Who developed opus-3?
- opus-3 was developed by Anthropic, based in the United States.
- When was opus-3 released?
- opus-3 was publicly released on March 4, 2024.
- How large is the context window of opus-3?
- AI IQ hasn't verified an input context limit for opus-3 from its developer's documentation, so none is listed rather than guessing.
- How fast is opus-3?
- AI IQ doesn't have independent speed measurements for opus-3 yet. Speed comes from independent tests of a model's hosted API, which aren't available for every model — especially new, restricted-access, or self-hosted ones.