Benchmark Almanac

Construction's Last Exam v2

Gemini 3.6 Flash 36.8% leads since October 2026
Not in model cards
Version 2

Models measure construction drawings, answer questions across sheets, and price quantities under contract rules. The board reports geometry, drawing reasoning, and estimating separately.

Togal reports scores for 22 models across the three tracks. The answer files contain scores on 60 tasks and nine additional unscored estimating tasks.

Made by
Togal.AI
Released
October 2026
Kind
Tasks from real jobs
Level
Expert
Scored
Checked automatically
Answers
Agent task
Measures
Capability
Language
English
In and out
Text, Documents → Text, Actions
Interaction
Agentic
Environment
APIs and tools
Items from
Real work and use
Run by
A host
In views
Agents
Strongest model tested
Claude Fable 5.1 #8 on the capability index
Last model added
September 2026
Scores refreshed
November 20th, 2026
9
Models scored on its own leaderboard

Gemini 3.6 Flash leads with 36.8% among models released by September 2026

Best score per model from the benchmark's own leaderboard, read on November 20th, 2026. Plain mean of the geometric, reasoning, and end-to-end track scores. 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 September 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 Gemini 3.6 Flash 36.8%
  2. 2 Claude Fable 5.1 35.3%
  3. 3 Gemini 3.7 Flash 31.6%
  4. 4 Claude Opus 5 31.5%
  5. 5 Gemini 3.5 Flash-Lite 22.4%
  6. 6 Claude Opus 4.6 21.6%
  7. 7 Gemini 3.1 Flash-Lite 19.1%
  8. 8 Claude Sonnet 5 14.3%
  9. 9 Claude Haiku 4.5 12.4%

Source: Official Construction's Last Exam v2 leaderboard

#ModelScoreSettingDateSource
1 Gemini 3.6 Flash 36.8% Jul 2026 Official leaderboard
2 Claude Fable 5.1 35.3% Sep 2026 Official leaderboard
3 Gemini 3.7 Flash 31.6% Aug 2026 Official leaderboard
4 Claude Opus 5 31.5% Jul 2026 Official leaderboard
5 Gemini 3.5 Flash-Lite 22.4% Jul 2026 Official leaderboard
6 Claude Opus 4.6 21.6% Feb 2026 Official leaderboard
7 Gemini 3.1 Flash-Lite 19.1% Mar 2026 Official leaderboard
8 Claude Sonnet 5 14.3% Jun 2026 Official leaderboard
9 Claude Haiku 4.5 12.4% Oct 2025 Official leaderboard

Rankings by part

We read 10 parts from the benchmark’s own leaderboard, each ranking models on one part of the test, such as Geometric, Reasoning and End-to-end. Gemini 3.7 Flash leads 3 of them, and 3 other models lead the rest. Arrows show how far a model moved from its overall rank among the models we track.

Geometric

9 models we track have a score here.

  1. 1 Gemini 3.7 Flash ↑2 19.2%
  2. 2 Gemini 3.6 Flash ↓1 13.8%
  3. 3 Claude Fable 5.1 ↓1 12.5%
  4. 4 Claude Opus 5 12.1%
  5. 5 Gemini 3.1 Flash-Lite ↑2 5.60%
  6. 6 Gemini 3.5 Flash-Lite ↓1 5.40%
  7. 7 Claude Haiku 4.5 ↑2 3.90%
  8. 8 Claude Opus 4.6 ↓2 3.80%
  9. 9 Claude Sonnet 5 ↓1 2.70%

The best score, 36.8%, costs $1.00 for the full run

Runs on the cost frontier score higher than any run that's cheaper. The costs are the leaderboard's own estimates, so compare them only within this benchmark.

Google DeepMind Cost frontier One model at each effort level
0% 10% 20% 30% 40% $0.10 $1.00 $10 $100 Cost of the full run, log scale Gemini 3.6 Flash Gemini 3.5 Flash-Lite Gemini 3.1 Flash-Lite 0% 10% 20% 30% 40% $0.10 $1.00 $10 $100 Cost of the full run, log scale Gemini 3.6 Flash Gemini 3.5 Flash-Lite

Source: Official Construction's Last Exam v2 leaderboard

ModelEffortHarnessScoreCost of the full runOutput tokens
Gemini 3.6 Flash 36.8% $1.00
Claude Fable 5.1 35.3% $13
Gemini 3.7 Flash 31.6% $1.10
Claude Opus 5 31.5% $7.50
Gemini 3.5 Flash-Lite 22.4% $0.60
Claude Opus 4.6 21.6% $6.50
Gemini 3.1 Flash-Lite 19.1% $0.40
Claude Sonnet 5 14.3% $3.00
Claude Haiku 4.5 12.4% $1.30

Number one over time

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

Google DeepMind
Own leaderboard Not tested since Oct 2026

Source: Official Construction's Last Exam v2 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%) Gemini 3.6 Flash Claude Opus 4.6 Claude Haiku 4.5 0% 20% 40% 60% 80% 100% 2026 Near the ceiling (85%) Gemini 3.6 Flash Claude Opus 4.6

Source: Official Construction's Last Exam v2 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.

Versions

We switch the top row to each new version when it comes out. Open it for one row per version. An old version stays in play while people still test it, so whoever first takes it past 85% gets credit for closing it out.

OpenAI Google DeepMind
Construction's Last Exam v2 Not tested since Oct 2026
Construction's Last Exam v2 Not tested since Oct 2026

Source: Official leaderboards

More in Economic work

All 210 in Economic work
  • JobBench In use 130 tasks from 35 occupations workers most want delegated. Agent works through messy reference files, graded by fact-anchored binary rubrics. GLM-5.3 61.4%
  • GDPval In use Real work tasks from 44 occupations with reference files. Experts blindly compare model deliverables with expert work. Metric is win-or-tie rate. Seed2.1 Pro 87.9%
  • GDPval-AA v2.1 In use Agents with shell and web access complete 220 GDPval gold tasks from 44 occupations. Blind pairwise comparisons give Elo ratings. Claude Opus 5.5 1,846
  • AA-Briefcase v1.1 In use Version 1.1 of Artificial Analysis's multi-hour office work evaluation, with Elo ratings fitted by Crowd-BT. Claude Opus 5.5 1,822
  • Agents' Last Exam v1.1 In use Agents complete 122 public long-horizon professional workflow tasks across 55 sub-industries. Deterministic graders check outcomes against hidden references. GPT-6.2 Astra 47.5%
  • CoWorkBench In use Agent completes long-horizon productivity tasks across computer science, finance, law and medicine. Qwen in-house scoring (method unpublished). Claude Fable 5 75.9%

Data audit

9 of 10 checks pass, 1 warning. Look first at the Task access check. We don’t know whether the tasks are public. Every benchmark

Catalog record 1 warning 1 pass
  • Warning Task access We don’t know whether the tasks are public.
  • Pass Confidence Checked against a primary source.
Scores 4 pass 1 note
  • Pass Independent runs Scores from the official leaderboard (9 entries).
  • Pass Models scored 9 models we track.
  • Pass Newest score The newest is from September 2026, for Claude Fable 5.1.
  • Note Frontier models 1 of the top 10 has a score, which is fine for a retired test. The strongest tested is Claude Fable 5.1 (#8).
  • Pass Headroom The best is 36.8%, under 85%, so it still separates models.
Leaderboard import 3 pass
  • Pass Importer We read the board on cle.togal.ai: 9 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