ArXivMath August 2026 with tools (ArXivMath Aug. 2026 (tools))
We show this table for reference; we do not rank on it.
The August 2026 ArXivMath release evaluated with a code-execution sandbox and no internet access.
Benchmark score on ArXivMath Aug. 2026 (tools) — September 27, 2026
We compile the ArXivMath Aug. 2026 (tools) rows from provider self-reports. Claude Opus 5.5 leads the table at 96.9%. We do not use these results to rank models overall.
1 modelMathematicsCurrentDisplay onlyUpdated September 27, 2026
Benchmark score table (1 model)
ScoreAbout ArXivMath Aug. 2026 (tools)
Year
2026
Tasks
57 recent research-mathematics problems
Format
Final-answer accuracy with tools
Difficulty
Research mathematics
Section 8.9 reports four-attempt average accuracy on the distinct 57-problem August 2026 release at max effort with tools. It remains separate from June and from MathArena's vendor-harness leaderboard.
Freshness and provenance
Version
ArXivMath Aug. 2026 (tools) 2026
Refresh cadence
Quarterly
Staleness state
Current
Question availability
Public benchmark set
BenchLM uses freshness metadata to decide whether a benchmark should still be treated as a strong differentiator, a benchmark to watch, or a display-only reference. For the full scoring policy, see the BenchLM methodology page.
Questions
What does ArXivMath Aug. 2026 (tools) measure?
The August 2026 ArXivMath release evaluated with a code-execution sandbox and no internet access.
Which model scores highest on ArXivMath Aug. 2026 (tools)?
Claude Opus 5.5 by Anthropic currently leads with a score of 96.9% on ArXivMath Aug. 2026 (tools).
How many models are evaluated on ArXivMath Aug. 2026 (tools)?
1 AI models have been evaluated on ArXivMath Aug. 2026 (tools) on BenchLM.
Know when it’s worth switching models
The model to choose, the cheaper alternative, and the release we would wait on.
Read a sample issueJoin 2,000+ readers.
One email each week. Unsubscribe anytime.