TerraBench: leaderboard
Metric: Hit@tol (%): share of answer fields within the field's tolerance band (exact match for strings and booleans), averaged over multi-field outputs, on the 403 executable TerraBench tasks (Fundamentals, Simulator-Grounded and Document-Grounded Verification tracks across eight Earth-science domains), each model driving the benchmark's TerraAgent ReAct harness with its 77 scientific tools; higher is better. Source: arxiv.org. Saturation forecast: Around 2028. 17 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Claude Sonnet 4.6 | 22.88 |
| 2 | GPT-5.5 | 21.03 |
| 3 | Claude Haiku 4.5 | 15.23 |
| 4 | Gemini 3.1 Pro (Preview) | 13.21 |
| 5 | GPT-5.4 | 9.8 |
| 6 | Gemini 2.5 Flash | 6.34 |
| 7 | Qwen 3.5 35B A3B | 5.89 |
| 8 | Qwen 3 14B | 3.8 |
| 9 | Gemma 4 26B A4B | 3.67 |
| 10 | Gemma 4 E4B | 2.9 |
| 11 | Mistral 7B Instruct (v0.3) | 1.67 |
| 12 | Qwen 3 8B | 1.33 |
| 13 | Qwen 3.5 9B | 1.2 |
| 14 | InternVL3-8B | 1.1 |
| 15 | Qwen 3 1.7B | 1.08 |
Interactive version: theaggregate.ai/benchmark?slug=terrabench · How It Works · Data refreshed daily, snapshot 2026-09-29.