Benchmark Almanac

Frontier Agentic ML & DS

Claude Opus 5 64.6% leads since August 2026
Not in model cards

Agents solve machine learning and data science tasks in a terminal. Verifiers check their deliverables and predictive quality.

The suite covers 304 tasks. Turing reports 7,296 graded runs.

Made by
Turing
Released
August 2026
Kind
Machine learning engineering
Level
Expert
Scored
Checked automatically
Answers
Agent task
Tasks
Kept private
Measures
Capability
Language
English
In and out
Text, Code → Code, Actions
Interaction
Agentic
Environment
Terminal
Refresh
One fixed set
Run by
A host
In views
Agents
Strongest model tested
Claude Opus 5 #11 on the capability index
Last model added
July 2026
Scores refreshed
November 20th, 2026
7
Models scored on its own leaderboard

Claude Opus 5 leads with 64.6% among models released by July 2026

Best score per model from the benchmark's own leaderboard, read on November 20th, 2026. Where a model ran with several settings or harnesses, we keep its best. Every run is in the next chart.

The newest model with a score here came out in July 2026, and nobody has tested a newer one. Among the 20 highest-ranked models on the capability index, these have no score yet:

  1. 1 Claude Opus 5 high 64.6%
  2. 2 GPT-5.6 Sol high 62.1%
  3. 3 GPT-5.5 medium 49.6%
  4. 4 Claude Opus 4.8 high 47%
  5. 5 Gemini 3.5 Flash medium 46.5%
  6. 6 Gemini 3.6 Flash medium 44.7%
  7. 7 GPT-5.6 Luna medium 30.6%

Source: Official Frontier Agentic ML & DS leaderboard

#ModelScore ±SettingDateSource
1 Claude Opus 5 64.6% ±0.9 high, Terminus-2 Jul 2026 Official leaderboard
2 GPT-5.6 Sol 62.1% ±1 high, Terminus-2 Jul 2026 Official leaderboard
3 GPT-5.5 49.6% ±1.1 medium, Terminus-2 Apr 2026 Official leaderboard
4 Claude Opus 4.8 47% ±1 high, Terminus-2 May 2026 Official leaderboard
5 Gemini 3.5 Flash 46.5% ±1.1 medium, Terminus-2 May 2026 Official leaderboard
6 Gemini 3.6 Flash 44.7% ±1 medium, Terminus-2 Jul 2026 Official leaderboard
7 GPT-5.6 Luna 30.6% ±1.1 medium, Terminus-2 Jul 2026 Official leaderboard

The best score, 64.6%, costs $1.37 a task

Runs on the cost frontier score higher than any run that's cheaper. GPT-5.6 Sol at medium comes within 5 points of the best for $0.35 a task, 4 times cheaper. The costs are the leaderboard's own estimates, so compare them only within this benchmark.

OpenAI Anthropic Cost frontier One model at each effort level
0% 20% 40% 60% 80% $0.01 $0.10 $1.00 $10 Cost per task, log scale Claude Opus 5, high GPT-5.6 Sol, medium GPT-5.6 Luna, medium 0% 20% 40% 60% 80% $0.01 $0.10 $1.00 $10 Cost per task, log scale Claude Opus 5, high GPT-5.6 Sol, medium

Source: Official Frontier Agentic ML & DS leaderboard

ModelEffortHarnessScoreCost per taskOutput tokens
Claude Opus 5 high Terminus-2 64.6% $1.37
GPT-5.6 Sol high Terminus-2 62.1% $0.60
GPT-5.6 Sol medium Terminus-2 59.6% $0.35
GPT-5.5 medium Terminus-2 49.6% $0.58
Claude Opus 4.8 high Terminus-2 47% $0.88
Gemini 3.5 Flash medium Terminus-2 46.5% $1.02
Gemini 3.6 Flash medium Terminus-2 44.7% $2.14
GPT-5.6 Luna medium Terminus-2 30.6% $0.02

Number one over time

Lab claims count from the day the card came out, independent runs from the model’s release.

Anthropic
Own leaderboard Not tested since Aug 2026

Source: Official Frontier Agentic ML & DS leaderboard

Scores over time

All the scores we have, one dot each, so a model tested at several settings appears more than once.

Plot with other benchmarks
On the benchmark’s own leaderboard Best so far
0% 20% 40% 60% 80% 100% 2026 Near the ceiling (85%) Claude Opus 5 GPT-5.5 0% 20% 40% 60% 80% 100% 2026 Near the ceiling (85%) Claude Opus 5 GPT-5.5

Source: Official Frontier Agentic ML & DS leaderboard

No lab model card reports it yet

We read the headline results of 540 launch posts from 57 labs, and none of them list this benchmark.

More in AI research

All 164 in AI research
  • PaperBench In use Agent replicates ICML 2024 papers from scratch. LLM judge grades against author-written hierarchical rubrics. Qwen 3.8 Max 93%
  • MLS-Bench In use Agent must invent one algorithmic change in an ML research scaffold whose gain transfers across settings. Scored relative to human baselines. Claude Fable 5.1 50.3%
  • PostTrainBench 1.2 In use CLI agent gets a base model, one H100 and a time budget to post-train it for a target benchmark. Scored on normalized gain. Claude Fable 5.1 44.6%
  • MLE-bench In use Kaggle-style ML engineering competitions. Agent trains models and is scored against competition leaderboards. GPT-6.2 Astra 96.1%
  • PMPP-Hard In use Agents write GPU kernels for 69 tasks that must be correct, and on 48 of them also run within 1.25 times the reference time. The score is the area under the solved-tasks curve over a 10-hour budget, out of 690 task-hours. Kimi K3 0.71
  • KernelBench In use Replace PyTorch operators with faster custom GPU kernels. Scored on correctness and speedup. MiniMax-M3 28.8%

Data audit

All 10 checks pass. Every benchmark

Catalog record 2 pass
  • Pass Task access The tasks are private.
  • Pass Confidence Checked against a primary source.
Scores 4 pass 1 note
  • Pass Independent runs Scores from the official leaderboard (7 entries).
  • Pass Models scored 7 models we track.
  • Pass Newest score The newest is from July 2026, for Claude Opus 5.
  • Note Frontier models None of the top 10 has a score. The strongest tested is Claude Opus 5 (#11).
  • Pass Headroom The best is 64.6%, under 85%, so it still separates models.
Leaderboard import 3 pass
  • Pass Importer We read the board on labs.turing.com: 8 rows for models we track at the last read.
  • Pass Last read November 20th, 2026.
  • Pass Read worked The last read worked.
Consistency 1 pass 1 note
  • Note Lab claims No model has both a lab number and an independent run.
  • Pass Known issues No published audit found problems.

How we check