METR Time Horizons
Integrity rank #42 of 65 · 34 models scored · top score 78.86 · Claude Opus 4.6
Autonomous agent capability as the length of task (in human completion time) an agent can complete at a given reliability (50%/80% time horizon).
Strongest on discrimination, weakest on contamination resistance. Caveat: scores come from an inconsistent mix of harnesses, so head-to-head comparisons are not apples-to-apples.
Reported scores — protocols may differ. How we compare
| # | Model | Score | Evidence | Measured |
|---|---|---|---|---|
| 1 | Claude Opus 4.6 | 78.86 | unverifiedT2 | 2026-02-05 |
| 2 | Gemini 3.1 Pro | 77.03 | self-reportedT1 | 2026-02-19 |
| 3 | GPT-5.2 | 75.29 | unverifiedT2 | 2025-12-11 |
| 4 | Claude Opus 4.5 | 74.97 | unverifiedT1 | 2025-11-24 |
| 5 | GPT-5.3 Codex | 74.54 | unverifiedT2 | 2026-02-05 |
| 6 | GPT-5.4 | 74.34 | self-reportedT1 | 2026-03-05 |
| 7 | Gemini 3 Pro | 70.98 | unverifiedT2 | 2025-11-18 |
| 8 | GPT-5 | 69.61 | unverifiedT1 | 2025-08-07 |
| 9 | Claude Sonnet 4.5 | 67.38 | unverifiedT1 | 2025-09-29 |
| 10 | Claude Opus 4.1 | 66.81 | unverifiedT1 | 2025-08-05 |
| 11 | Grok 4 | 66.58 | unverifiedT1 | 2025-07-09 |
| 12 | o3 | 65.44 | unverifiedT1 | 2025-04-16 |
| 13 | Claude Opus 4 | 63.93 | unverifiedT1 | 2025-05-22 |
| 14 | o4-mini | 63.93 | unverifiedT1 | 2025-04-16 |
| 15 | Claude Sonnet 4 | 61.97 | unverifiedT1 | 2025-05-22 |
| 16 | Claude 3.7 Sonnet | 59.97 | unverifiedT1 | 2025-02-24 |
| 17 | Kimi K2 Thinking | 59.19 | unverifiedT1 | 2025-11-06 |
| 18 | gpt-oss-120b | 56.63 | unverifiedT1 | 2025-08-05 |
| 19 | Gemini 2.5 Pro (Jun 2025) | 55.44 | unverifiedT1 | 2025-06-05 |
| 20 | DeepSeek-R1 (May 2025) | 53.78 | unverifiedT1 | 2025-05-28 |
| 21 | DeepSeek-R1 | 51.93 | unverifiedT1 | 2025-01-20 |
| 22 | o1 | 51.05 | self-reportedT1 | 2024-12-17 |
| 23 | DeepSeek-V3 (Mar 2025) | 49.58 | unverifiedT1 | 2025-03-24 |
| 24 | DeepSeek-V3 | 47.36 | unverifiedT1 | 2024-12-26 |
| 25 | Claude 3.5 Sonnet (October 2024) | 45.24 | self-reportedT1 | 2024-10-22 |
| 26 | o1-preview | 45.14 | self-reportedT1 | 2024-09-12 |
| 27 | GPT-4o (May 2024) | 40.76 | unverifiedT1 | 2024-05-13 |
| 28 | Claude 3.5 Sonnet | 40.15 | self-reportedT1 | 2024-06-20 |
| 29 | GPT-4 Turbo (Apr 2024) | 36.74 | unverifiedT1 | 2024-04-09 |
| 30 | GPT-4 (Mar 2023) | 36.11 | unverifiedT1 | 2023-03-14 |
| 31 | Qwen2.5-72B | 35.78 | unverifiedT1 | 2024-09-19 |
| 32 | GPT-4 Turbo (Nov 2023) | 35.17 | unverifiedT1 | 2024-01-25 |
| 33 | Qwen2-72B | 29.9 | unverifiedT1 | 2024-06-07 |
| 34 | Claude 3 Opus | 29.48 | self-reportedT1 | 2024-02-29 |
Continue from this leaderboard into the tightest head-to-head reads.
Weighted composite. Each dimension opens to the source-backed sub-signals it is computed from.
Claims drawn from cited facts, not live model generation.
This benchmark measures the duration of tasks an autonomous agent can complete at specified reliability levels, using human completion time.
1 cited fact
This benchmark has scores from 34 tracked models, with a recorded score range spanning 49.38.
2 cited facts
Its Benchmark Integrity Index score is 81, with grade B and rank 42 out of 65 tracked benchmarks.
4 cited facts
It is not flagged as saturated. The top recorded score is 78.86, with 21.14 points of headroom. The computed top-score cluster contains 2 models.
4 cited facts
The harness classification is mixed harness — not directly comparable. Test-set privacy is recorded as unknown. The recorded contamination-window ratio is 0.588.
3 cited facts
METR Time Horizons: METR Time Horizons as reported in Epoch AI's Capabilities Index CSV.
Claude Opus 4.6 leads METR Time Horizons at 78.86. The full leaderboard above lists every recorded measurement, not just the headline number.
34 models have recorded scores on METR Time Horizons, spanning a score spread of 49.38.
tensor.news grades METR Time Horizons B for integrity (score 81/100), ranking #42 of 65 benchmarks we assess across discrimination, saturation, contamination resistance, harness comparability, and freshness.
No — METR Time Horizons still has headroom and continues to discriminate between models rather than bunching them at the ceiling.
Its test set is unknown, so contamination risk is on the table: a high score may partly reflect training-data overlap rather than capability.
Only partly — scores come from a mix of harnesses and protocols, so a head-to-head on METR Time Horizons is not fully apples-to-apples.
Every METR Time Horizons measurement is source-backed and tagged reproduced, self-reported, or unverified; the underlying record cites metr.org/time-horizons.