FEM-Bench
Models implement finite element functions and write tests for them. The evaluator checks numerical outputs and whether the tests distinguish reference functions from faulty implementations.
The authors compare ten models on 33 tasks from computational mechanics coursework. They publish separate results for function implementation and test generation.
- Made by
- Lejeune Lab at Boston University and collaborators
- Released
- August 2025
- Kind
- Scientific code
- Level
- University
- Scored
- Checked automatically
- Answers
- Code
- Tasks
- All public
- Measures
- Capability
- Language
- English
- Interaction
- Single turn
- Items from
- Exams and contests
- Refresh
- One fixed set
- Run by
- Anyone
- In views
- CAD
- Strongest model tested
-
Gemini 3 Pro #56 on the capability index
- Last model added
- November 2025
- Scores refreshed
- November 20th, 2026
Gemini 3 Pro leads with 87.9% among models released by November 2025 Gemini 3 Pro leads with 87.9% among models released by November 2025
Best score per entry from the benchmark's own leaderboard, read on November 20th, 2026. Where a model ran with several settings or harnesses, we keep its best. Names in gray aren't in our model list: fine-tunes, agent systems, products and older models. They don't count toward leaders. This view shows only models we track, each at its best.
The newest model with a score here came out in November 2025, and nobody has tested a newer one. Among the 20 highest-ranked models on the capability index, these have no score yet:
- GPT-6.2 Astra
- Claude Fable 5.2
-
Gemini 4 Carbon
- GPT-6 Astra
- and 16 more
- 1
Gemini 3 Pro First run 87.9%
- 2 Claude Opus 4.5 First run 84.8%
- 3
Gemini 2.5 Pro Jun 2025, First run 78.8%
- 4 GPT-5 mini First run 75.8%
- 5 GPT-5 First run 66.7%
- 6
Qwen3-Coder-480B-A35B First run 63.6%
- 7 Claude Haiku 4.5 First run 57.6%
- 8 qwen3-next-80b First run 48.5%
- 9
Llama 4 Maverick First run 48.5%
- 10
Llama 4 Scout First run 18.2%
Source: Official FEM-Bench leaderboard
- 1
Gemini 3 Pro First run 87.9%
- 2 Claude Opus 4.5 First run 84.8%
- 3
Gemini 2.5 Pro Jun 2025, First run 78.8%
- 4 GPT-5 mini First run 75.8%
- 5 GPT-5 First run 66.7%
- 6
Qwen3-Coder-480B-A35B First run 63.6%
- 7 Claude Haiku 4.5 First run 57.6%
- 8
Llama 4 Maverick First run 48.5%
- 9
Llama 4 Scout First run 18.2%
Source: Official FEM-Bench leaderboard
| # | Entry | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | | 87.9% | First run | Nov 2025 | Official leaderboard |
| 2 | Claude Opus 4.5 | 84.8% | First run | Nov 2025 | Official leaderboard |
| 3 | | 78.8% | First run | Jun 2025 | Official leaderboard |
| 4 | GPT-5 mini | 75.8% | First run | Aug 2025 | Official leaderboard |
| 5 | GPT-5 | 66.7% | First run | Aug 2025 | Official leaderboard |
| 6 | | 63.6% | First run | Jul 2025 | Official leaderboard |
| 7 | Claude Haiku 4.5 | 57.6% | First run | Oct 2025 | Official leaderboard |
| 8 | qwen3-next-80b | 48.5% | First run | Official leaderboard | |
| 9 | | 48.5% | First run | Apr 2025 | Official leaderboard |
| 10 | | 18.2% | First run | Apr 2025 | Official leaderboard |
| # | Model | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | | 87.9% | First run | Nov 2025 | Official leaderboard |
| 2 | Claude Opus 4.5 | 84.8% | First run | Nov 2025 | Official leaderboard |
| 3 | | 78.8% | First run | Jun 2025 | Official leaderboard |
| 4 | GPT-5 mini | 75.8% | First run | Aug 2025 | Official leaderboard |
| 5 | GPT-5 | 66.7% | First run | Aug 2025 | Official leaderboard |
| 6 | | 63.6% | First run | Jul 2025 | Official leaderboard |
| 7 | Claude Haiku 4.5 | 57.6% | First run | Oct 2025 | Official leaderboard |
| 8 | | 48.5% | First run | Apr 2025 | Official leaderboard |
| 9 | | 18.2% | First run | Apr 2025 | Official leaderboard |
Rankings by part
We read 66 parts from the benchmark’s own leaderboard in 2 groups: function correctness and joint test success. Each ranks models on one part of the test. Llama 4 Maverick leads 31 of them, and 7 other models lead the rest. Arrows show how far a model moved from its overall rank among the models we track.
linear_elastic_CC0_H0_T0
9 models we track have a score here.
- 1
Llama 4 Maverick ↑7 100%
- 2
Gemini 2.5 Pro Jun 2025 ↑1 100%
- 3
Qwen3-Coder-480B-A35B ↑3 100%
- 4 GPT-5 ↑1 100%
- 5 GPT-5 mini ↓1 100%
- 6
Gemini 3 Pro ↓5 100%
- 7 Claude Opus 4.5 ↓5 100%
- 8
Llama 4 Scout ↑1 0%
- 9 Claude Haiku 4.5 ↓2 0%
local_elastic_stiffness_CC0_H3_T1
9 models we track have a score here.
- 1
Llama 4 Scout ↑8 100%
- 2
Gemini 2.5 Pro Jun 2025 ↑1 100%
- 3
Qwen3-Coder-480B-A35B ↑3 100%
- 4
Gemini 3 Pro ↓3 100%
- 5 Claude Opus 4.5 ↓3 100%
- 6
Llama 4 Maverick ↑2 0%
- 7 GPT-5 ↓2 0%
- 8 GPT-5 mini ↓4 0%
- 9 Claude Haiku 4.5 ↓2 0%
uniform_mesh_CC0_H0_T0
9 models we track have a score here.
- 1
Llama 4 Maverick ↑7 100%
- 2
Llama 4 Scout ↑7 100%
- 3
Gemini 2.5 Pro Jun 2025 100%
- 4
Qwen3-Coder-480B-A35B ↑2 100%
- 5 GPT-5 mini ↓1 100%
- 6 Claude Haiku 4.5 ↑1 100%
- 7
Gemini 3 Pro ↓6 100%
- 8 Claude Opus 4.5 ↓6 100%
- 9 GPT-5 ↓4 0%
quad8_element_distributed_load_CC0_H0_T0
9 models we track have a score here.
- 1
Gemini 2.5 Pro Jun 2025 ↑2 100%
- 2
Qwen3-Coder-480B-A35B ↑4 100%
- 3 GPT-5 ↑2 100%
- 4 GPT-5 mini 100%
- 5 Claude Haiku 4.5 ↑2 100%
- 6
Gemini 3 Pro ↓5 100%
- 7 Claude Opus 4.5 ↓5 100%
- 8
Llama 4 Maverick 0%
- 9
Llama 4 Scout 0%
quad8_integral_of_derivative_CC0_H3_T3
9 models we track have a score here.
- 1
Gemini 2.5 Pro Jun 2025 ↑2 100%
- 2 GPT-5 mini ↑2 100%
- 3
Gemini 3 Pro ↓2 100%
- 4
Llama 4 Maverick ↑4 0%
- 5
Llama 4 Scout ↑4 0%
- 6
Qwen3-Coder-480B-A35B 0%
- 7 GPT-5 ↓2 0%
- 8 Claude Haiku 4.5 ↓1 0%
- 9 Claude Opus 4.5 ↓7 0%
quad8_mesh_rectangle_CC0_H0_T0
9 models we track have a score here.
- 1
Gemini 2.5 Pro Jun 2025 ↑2 100%
- 2
Qwen3-Coder-480B-A35B ↑4 100%
- 3 GPT-5 ↑2 100%
- 4 GPT-5 mini 100%
- 5 Claude Haiku 4.5 ↑2 100%
- 6
Gemini 3 Pro ↓5 100%
- 7 Claude Opus 4.5 ↓5 100%
- 8
Llama 4 Maverick 0%
- 9
Llama 4 Scout 0%
quad8_physical_gradient_CC0_H1_T3
9 models we track have a score here.
- 1
Gemini 2.5 Pro Jun 2025 ↑2 100%
- 2 GPT-5 ↑3 100%
- 3 GPT-5 mini ↑1 100%
- 4
Gemini 3 Pro ↓3 100%
- 5
Llama 4 Maverick ↑3 0%
- 6
Llama 4 Scout ↑3 0%
- 7
Qwen3-Coder-480B-A35B ↓1 0%
- 8 Claude Haiku 4.5 ↓1 0%
- 9 Claude Opus 4.5 ↓7 0%
quad8_shape_fcns_and_derivatives_CC0_H0_T0
9 models we track have a score here.
- 1
Gemini 2.5 Pro Jun 2025 ↑2 100%
- 2 GPT-5 ↑3 100%
- 3 GPT-5 mini ↑1 100%
- 4
Gemini 3 Pro ↓3 100%
- 5 Claude Opus 4.5 ↓3 100%
- 6
Llama 4 Maverick ↑2 0%
- 7
Llama 4 Scout ↑2 0%
- 8
Qwen3-Coder-480B-A35B ↓2 0%
- 9 Claude Haiku 4.5 ↓2 0%
quad_quadrature_CC0_H0_T0
9 models we track have a score here.
- 1
Llama 4 Maverick ↑7 100%
- 2
Gemini 2.5 Pro Jun 2025 ↑1 100%
- 3
Qwen3-Coder-480B-A35B ↑3 100%
- 4 GPT-5 ↑1 100%
- 5 GPT-5 mini ↓1 100%
- 6 Claude Haiku 4.5 ↑1 100%
- 7
Gemini 3 Pro ↓6 100%
- 8 Claude Opus 4.5 ↓6 100%
- 9
Llama 4 Scout 0%
tri6_mesh_rectangle_CC0_H0_T0
9 models we track have a score here.
- 1
Gemini 2.5 Pro Jun 2025 ↑2 100%
- 2
Qwen3-Coder-480B-A35B ↑4 100%
- 3 GPT-5 ↑2 100%
- 4 GPT-5 mini 100%
- 5 Claude Haiku 4.5 ↑2 100%
- 6
Gemini 3 Pro ↓5 100%
- 7 Claude Opus 4.5 ↓5 100%
- 8
Llama 4 Maverick 0%
- 9
Llama 4 Scout 0%
tri6_shape_fcns_and_derivatives_CC0_H0_T0
9 models we track have a score here.
- 1
Llama 4 Maverick ↑7 100%
- 2
Gemini 2.5 Pro Jun 2025 ↑1 100%
- 3
Qwen3-Coder-480B-A35B ↑3 100%
- 4 GPT-5 ↑1 100%
- 5 GPT-5 mini ↓1 100%
- 6 Claude Haiku 4.5 ↑1 100%
- 7
Gemini 3 Pro ↓6 100%
- 8 Claude Opus 4.5 ↓6 100%
- 9
Llama 4 Scout 0%
tri_quadrature_CC0_H0_T0
9 models we track have a score here.
- 1
Llama 4 Maverick ↑7 100%
- 2
Llama 4 Scout ↑7 100%
- 3
Gemini 2.5 Pro Jun 2025 100%
- 4
Qwen3-Coder-480B-A35B ↑2 100%
- 5 GPT-5 100%
- 6 GPT-5 mini ↓2 100%
- 7 Claude Haiku 4.5 100%
- 8
Gemini 3 Pro ↓7 100%
- 9 Claude Opus 4.5 ↓7 100%
assemble_global_geometric_stiffness_CC1_H4_T1
9 models we track have a score here.
- 1
Llama 4 Maverick ↑7 100%
- 2
Llama 4 Scout ↑7 100%
- 3
Gemini 2.5 Pro Jun 2025 100%
- 4
Qwen3-Coder-480B-A35B ↑2 100%
- 5 GPT-5 100%
- 6 GPT-5 mini ↓2 100%
- 7 Claude Haiku 4.5 100%
- 8
Gemini 3 Pro ↓7 100%
- 9 Claude Opus 4.5 ↓7 100%
assemble_global_geometric_stiffness_CC1_H4_T2
9 models we track have a score here.
- 1 Claude Opus 4.5 ↑1 100%
- 2
Llama 4 Maverick ↑6 0%
- 3
Llama 4 Scout ↑6 0%
- 4
Gemini 2.5 Pro Jun 2025 ↓1 0%
- 5
Qwen3-Coder-480B-A35B ↑1 0%
- 6 GPT-5 ↓1 0%
- 7 GPT-5 mini ↓3 0%
- 8 Claude Haiku 4.5 ↓1 0%
- 9
Gemini 3 Pro ↓8 0%
assemble_global_geometric_stiffness_CC1_H4_T3
9 models we track have a score here.
- 1
Llama 4 Maverick ↑7 0%
- 2
Llama 4 Scout ↑7 0%
- 3
Gemini 2.5 Pro Jun 2025 0%
- 4
Qwen3-Coder-480B-A35B ↑2 0%
- 5 GPT-5 0%
- 6 GPT-5 mini ↓2 0%
- 7 Claude Haiku 4.5 0%
- 8
Gemini 3 Pro ↓7 0%
- 9 Claude Opus 4.5 ↓7 0%
assemble_global_linear_elastic_stiffness_CC0_H2_T1
9 models we track have a score here.
- 1
Llama 4 Maverick ↑7 100%
- 2
Gemini 2.5 Pro Jun 2025 ↑1 100%
- 3
Qwen3-Coder-480B-A35B ↑3 100%
- 4 GPT-5 ↑1 100%
- 5 GPT-5 mini ↓1 100%
- 6 Claude Haiku 4.5 ↑1 100%
- 7
Gemini 3 Pro ↓6 100%
- 8 Claude Opus 4.5 ↓6 100%
- 9
Llama 4 Scout 0%
assemble_global_linear_elastic_stiffness_CC0_H2_T3
9 models we track have a score here.
- 1 GPT-5 ↑4 100%
- 2
Gemini 3 Pro ↓1 100%
- 3 Claude Opus 4.5 ↓1 100%
- 4
Llama 4 Maverick ↑4 0%
- 5
Llama 4 Scout ↑4 0%
- 6
Gemini 2.5 Pro Jun 2025 ↓3 0%
- 7
Qwen3-Coder-480B-A35B ↓1 0%
- 8 GPT-5 mini ↓4 0%
- 9 Claude Haiku 4.5 ↓2 0%
assemble_global_load_CC0_H0_T0
9 models we track have a score here.
- 1
Llama 4 Maverick ↑7 100%
- 2
Gemini 2.5 Pro Jun 2025 ↑1 100%
- 3
Qwen3-Coder-480B-A35B ↑3 100%
- 4 GPT-5 ↑1 100%
- 5 GPT-5 mini ↓1 100%
- 6 Claude Haiku 4.5 ↑1 100%
- 7
Gemini 3 Pro ↓6 100%
- 8 Claude Opus 4.5 ↓6 100%
- 9
Llama 4 Scout 0%
elastic_critical_load_CC1_H10_T1
9 models we track have a score here.
- 1
Llama 4 Maverick ↑7 100%
- 2
Gemini 2.5 Pro Jun 2025 ↑1 100%
- 3
Qwen3-Coder-480B-A35B ↑3 100%
- 4 GPT-5 ↑1 100%
- 5 GPT-5 mini ↓1 100%
- 6 Claude Haiku 4.5 ↑1 100%
- 7
Gemini 3 Pro ↓6 100%
- 8 Claude Opus 4.5 ↓6 100%
- 9
Llama 4 Scout 0%
elastic_critical_load_CC1_H10_T2
9 models we track have a score here.
- 1
Gemini 3 Pro 100%
- 2 Claude Opus 4.5 100%
- 3
Llama 4 Maverick ↑5 0%
- 4
Llama 4 Scout ↑5 0%
- 5
Gemini 2.5 Pro Jun 2025 ↓2 0%
- 6
Qwen3-Coder-480B-A35B 0%
- 7 GPT-5 ↓2 0%
- 8 GPT-5 mini ↓4 0%
- 9 Claude Haiku 4.5 ↓2 0%
elastic_critical_load_CC1_H10_T3
9 models we track have a score here.
- 1
Llama 4 Maverick ↑7 0%
- 2
Llama 4 Scout ↑7 0%
- 3
Gemini 2.5 Pro Jun 2025 0%
- 4
Qwen3-Coder-480B-A35B ↑2 0%
- 5 GPT-5 0%
- 6 GPT-5 mini ↓2 0%
- 7 Claude Haiku 4.5 0%
- 8
Gemini 3 Pro ↓7 0%
- 9 Claude Opus 4.5 ↓7 0%
linear_elastic_CC0_H6_T1
9 models we track have a score here.
- 1
Llama 4 Maverick ↑7 100%
- 2
Llama 4 Scout ↑7 100%
- 3
Gemini 2.5 Pro Jun 2025 100%
- 4
Qwen3-Coder-480B-A35B ↑2 100%
- 5 GPT-5 100%
- 6 GPT-5 mini ↓2 100%
- 7 Claude Haiku 4.5 100%
- 8
Gemini 3 Pro ↓7 100%
- 9 Claude Opus 4.5 ↓7 100%
linear_elastic_CC0_H6_T3
9 models we track have a score here.
- 1
Gemini 2.5 Pro Jun 2025 ↑2 100%
- 2 GPT-5 mini ↑2 100%
- 3
Gemini 3 Pro ↓2 100%
- 4 Claude Opus 4.5 ↓2 100%
- 5
Llama 4 Maverick ↑3 0%
- 6
Llama 4 Scout ↑3 0%
- 7
Qwen3-Coder-480B-A35B ↓1 0%
- 8 GPT-5 ↓3 0%
- 9 Claude Haiku 4.5 ↓2 0%
local_elastic_stiffness_CC0_H0_T0
9 models we track have a score here.
- 1
Llama 4 Maverick ↑7 100%
- 2
Gemini 2.5 Pro Jun 2025 ↑1 100%
- 3
Qwen3-Coder-480B-A35B ↑3 100%
- 4 Claude Haiku 4.5 ↑3 100%
- 5
Gemini 3 Pro ↓4 100%
- 6 Claude Opus 4.5 ↓4 100%
- 7
Llama 4 Scout ↑2 0%
- 8 GPT-5 ↓3 0%
- 9 GPT-5 mini ↓5 0%
local_element_loads_CC0_H2_T1
9 models we track have a score here.
- 1
Llama 4 Maverick ↑7 100%
- 2
Gemini 2.5 Pro Jun 2025 ↑1 100%
- 3
Qwen3-Coder-480B-A35B ↑3 100%
- 4 GPT-5 ↑1 100%
- 5 GPT-5 mini ↓1 100%
- 6 Claude Haiku 4.5 ↑1 100%
- 7
Gemini 3 Pro ↓6 100%
- 8 Claude Opus 4.5 ↓6 100%
- 9
Llama 4 Scout 0%
local_element_loads_CC0_H2_T3
9 models we track have a score here.
- 1
Gemini 2.5 Pro Jun 2025 ↑2 100%
- 2 GPT-5 ↑3 100%
- 3 GPT-5 mini ↑1 100%
- 4
Gemini 3 Pro ↓3 100%
- 5 Claude Opus 4.5 ↓3 100%
- 6
Llama 4 Maverick ↑2 0%
- 7
Llama 4 Scout ↑2 0%
- 8
Qwen3-Coder-480B-A35B ↓2 0%
- 9 Claude Haiku 4.5 ↓2 0%
local_geometric_stiffness_CC1_H0_T0
9 models we track have a score here.
- 1
Llama 4 Maverick ↑7 0%
- 2
Llama 4 Scout ↑7 0%
- 3
Gemini 2.5 Pro Jun 2025 0%
- 4
Qwen3-Coder-480B-A35B ↑2 0%
- 5 GPT-5 0%
- 6 GPT-5 mini ↓2 0%
- 7 Claude Haiku 4.5 0%
- 8
Gemini 3 Pro ↓7 0%
- 9 Claude Opus 4.5 ↓7 0%
partition_DOFs_CC0_H0_T0
9 models we track have a score here.
- 1
Llama 4 Maverick ↑7 100%
- 2
Gemini 2.5 Pro Jun 2025 ↑1 100%
- 3
Qwen3-Coder-480B-A35B ↑3 100%
- 4 GPT-5 ↑1 100%
- 5 GPT-5 mini ↓1 100%
- 6 Claude Haiku 4.5 ↑1 100%
- 7
Gemini 3 Pro ↓6 100%
- 8 Claude Opus 4.5 ↓6 100%
- 9
Llama 4 Scout 0%
solve_eigenvalue_CC1_H1_T1
9 models we track have a score here.
- 1
Llama 4 Maverick ↑7 100%
- 2
Gemini 2.5 Pro Jun 2025 ↑1 100%
- 3
Qwen3-Coder-480B-A35B ↑3 100%
- 4 GPT-5 ↑1 100%
- 5 GPT-5 mini ↓1 100%
- 6
Gemini 3 Pro ↓5 100%
- 7 Claude Opus 4.5 ↓5 100%
- 8
Llama 4 Scout ↑1 0%
- 9 Claude Haiku 4.5 ↓2 0%
solve_eigenvalue_CC1_H1_T3
9 models we track have a score here.
- 1
Llama 4 Maverick ↑7 100%
- 2
Gemini 2.5 Pro Jun 2025 ↑1 100%
- 3
Qwen3-Coder-480B-A35B ↑3 100%
- 4 GPT-5 ↑1 100%
- 5 GPT-5 mini ↓1 100%
- 6 Claude Haiku 4.5 ↑1 100%
- 7
Gemini 3 Pro ↓6 100%
- 8 Claude Opus 4.5 ↓6 100%
- 9
Llama 4 Scout 0%
solve_linear_CC0_H1_T1
9 models we track have a score here.
- 1
Gemini 2.5 Pro Jun 2025 ↑2 100%
- 2
Qwen3-Coder-480B-A35B ↑4 100%
- 3 GPT-5 ↑2 100%
- 4 GPT-5 mini 100%
- 5 Claude Haiku 4.5 ↑2 100%
- 6
Gemini 3 Pro ↓5 100%
- 7 Claude Opus 4.5 ↓5 100%
- 8
Llama 4 Maverick 0%
- 9
Llama 4 Scout 0%
solve_linear_CC0_H1_T3
9 models we track have a score here.
- 1
Llama 4 Maverick ↑7 100%
- 2
Llama 4 Scout ↑7 100%
- 3
Gemini 2.5 Pro Jun 2025 100%
- 4
Qwen3-Coder-480B-A35B ↑2 100%
- 5 GPT-5 100%
- 6 GPT-5 mini ↓2 100%
- 7 Claude Haiku 4.5 100%
- 8
Gemini 3 Pro ↓7 100%
- 9 Claude Opus 4.5 ↓7 100%
transformation_matrix_CC0_H0_T0
9 models we track have a score here.
- 1 GPT-5 mini ↑3 100%
- 2 Claude Haiku 4.5 ↑5 100%
- 3
Gemini 3 Pro ↓2 100%
- 4 Claude Opus 4.5 ↓2 100%
- 5
Llama 4 Maverick ↑3 0%
- 6
Llama 4 Scout ↑3 0%
- 7
Gemini 2.5 Pro Jun 2025 ↓4 0%
- 8
Qwen3-Coder-480B-A35B ↓2 0%
- 9 GPT-5 ↓4 0%
linear_elastic_CC0_H0_T0
8 models we track have a score here.
- 1
Llama 4 Scout ↑8 100%
- 2
Gemini 2.5 Pro Jun 2025 ↑1 100%
- 3 GPT-5 ↑2 100%
- 4 GPT-5 mini 100%
- 5 Claude Haiku 4.5 ↑2 100%
- 6
Gemini 3 Pro ↓5 100%
- 7 Claude Opus 4.5 ↓5 100%
- 8
Qwen3-Coder-480B-A35B ↓2 50%
local_elastic_stiffness_CC0_H3_T1
9 models we track have a score here.
- 1
Llama 4 Maverick ↑7 100%
- 2
Llama 4 Scout ↑7 100%
- 3
Gemini 2.5 Pro Jun 2025 100%
- 4
Qwen3-Coder-480B-A35B ↑2 100%
- 5 GPT-5 mini ↓1 100%
- 6 Claude Haiku 4.5 ↑1 100%
- 7
Gemini 3 Pro ↓6 100%
- 8 Claude Opus 4.5 ↓6 100%
- 9 GPT-5 ↓4 0%
uniform_mesh_CC0_H0_T0
9 models we track have a score here.
- 1
Llama 4 Maverick ↑7 100%
- 2
Llama 4 Scout ↑7 100%
- 3
Gemini 2.5 Pro Jun 2025 100%
- 4
Qwen3-Coder-480B-A35B ↑2 100%
- 5 GPT-5 100%
- 6 GPT-5 mini ↓2 100%
- 7 Claude Haiku 4.5 100%
- 8
Gemini 3 Pro ↓7 100%
- 9 Claude Opus 4.5 ↓7 100%
quad8_element_distributed_load_CC0_H0_T0
9 models we track have a score here.
- 1
Gemini 2.5 Pro Jun 2025 ↑2 100%
- 2 GPT-5 ↑3 100%
- 3 Claude Haiku 4.5 ↑4 100%
- 4
Gemini 3 Pro ↓3 100%
- 5 Claude Opus 4.5 ↓3 100%
- 6
Qwen3-Coder-480B-A35B 50%
- 7 GPT-5 mini ↓3 50%
- 8
Llama 4 Maverick 0%
- 9
Llama 4 Scout 0%
quad8_integral_of_derivative_CC0_H3_T3
9 models we track have a score here.
- 1 GPT-5 ↑4 100%
- 2 Claude Opus 4.5 100%
- 3
Gemini 2.5 Pro Jun 2025 66.7%
- 4
Gemini 3 Pro ↓3 66.7%
- 5
Llama 4 Maverick ↑3 33.3%
- 6 GPT-5 mini ↓2 33.3%
- 7 Claude Haiku 4.5 33.3%
- 8
Llama 4 Scout ↑1 0%
- 9
Qwen3-Coder-480B-A35B ↓3 0%
quad8_mesh_rectangle_CC0_H0_T0
9 models we track have a score here.
- 1 GPT-5 ↑4 100%
- 2
Gemini 3 Pro ↓1 100%
- 3
Llama 4 Maverick ↑5 66.7%
- 4
Gemini 2.5 Pro Jun 2025 ↓1 66.7%
- 5
Qwen3-Coder-480B-A35B ↑1 66.7%
- 6 GPT-5 mini ↓2 66.7%
- 7 Claude Haiku 4.5 66.7%
- 8 Claude Opus 4.5 ↓6 66.7%
- 9
Llama 4 Scout 33.3%
quad8_physical_gradient_CC0_H1_T3
9 models we track have a score here.
- 1
Llama 4 Maverick ↑7 100%
- 2
Gemini 2.5 Pro Jun 2025 ↑1 100%
- 3 GPT-5 ↑2 100%
- 4 GPT-5 mini 100%
- 5 Claude Haiku 4.5 ↑2 100%
- 6
Gemini 3 Pro ↓5 100%
- 7 Claude Opus 4.5 ↓5 100%
- 8
Llama 4 Scout ↑1 50%
- 9
Qwen3-Coder-480B-A35B ↓3 50%
quad8_shape_fcns_and_derivatives_CC0_H0_T0
9 models we track have a score here.
- 1 GPT-5 ↑4 100%
- 2 GPT-5 mini ↑2 100%
- 3 Claude Haiku 4.5 ↑4 100%
- 4
Gemini 3 Pro ↓3 100%
- 5 Claude Opus 4.5 ↓3 100%
- 6
Llama 4 Maverick ↑2 83.3%
- 7
Gemini 2.5 Pro Jun 2025 ↓4 83.3%
- 8
Qwen3-Coder-480B-A35B ↓2 83.3%
- 9
Llama 4 Scout 50%
quad_quadrature_CC0_H0_T0
9 models we track have a score here.
- 1
Llama 4 Scout ↑8 100%
- 2
Gemini 2.5 Pro Jun 2025 ↑1 100%
- 3
Qwen3-Coder-480B-A35B ↑3 100%
- 4 GPT-5 ↑1 100%
- 5 GPT-5 mini ↓1 100%
- 6 Claude Haiku 4.5 ↑1 100%
- 7 Claude Opus 4.5 ↓5 100%
- 8
Llama 4 Maverick 60%
- 9
Gemini 3 Pro ↓8 40%
tri6_mesh_rectangle_CC0_H0_T0
9 models we track have a score here.
- 1
Llama 4 Maverick ↑7 100%
- 2
Gemini 2.5 Pro Jun 2025 ↑1 100%
- 3
Qwen3-Coder-480B-A35B ↑3 100%
- 4 GPT-5 ↑1 100%
- 5 GPT-5 mini ↓1 100%
- 6 Claude Haiku 4.5 ↑1 100%
- 7 Claude Opus 4.5 ↓5 100%
- 8
Llama 4 Scout ↑1 66.7%
- 9
Gemini 3 Pro ↓8 66.7%
tri6_shape_fcns_and_derivatives_CC0_H0_T0
9 models we track have a score here.
- 1
Gemini 2.5 Pro Jun 2025 ↑2 100%
- 2 GPT-5 mini ↑2 100%
- 3
Gemini 3 Pro ↓2 100%
- 4 Claude Opus 4.5 ↓2 100%
- 5 GPT-5 83.3%
- 6
Llama 4 Maverick ↑2 50%
- 7
Qwen3-Coder-480B-A35B ↓1 50%
- 8 Claude Haiku 4.5 ↓1 50%
- 9
Llama 4 Scout 33.3%
tri_quadrature_CC0_H0_T0
9 models we track have a score here.
- 1
Qwen3-Coder-480B-A35B ↑5 100%
- 2 GPT-5 mini ↑2 100%
- 3 Claude Haiku 4.5 ↑4 100%
- 4
Llama 4 Maverick ↑4 60%
- 5
Llama 4 Scout ↑4 60%
- 6 GPT-5 ↓1 60%
- 7
Gemini 2.5 Pro Jun 2025 ↓4 40%
- 8
Gemini 3 Pro ↓7 40%
- 9 Claude Opus 4.5 ↓7 40%
assemble_global_geometric_stiffness_CC1_H4_T1
9 models we track have a score here.
- 1 GPT-5 ↑4 100%
- 2
Gemini 2.5 Pro Jun 2025 ↑1 50%
- 3
Qwen3-Coder-480B-A35B ↑3 50%
- 4
Llama 4 Maverick ↑4 0%
- 5
Llama 4 Scout ↑4 0%
- 6 GPT-5 mini ↓2 0%
- 7 Claude Haiku 4.5 0%
- 8
Gemini 3 Pro ↓7 0%
- 9 Claude Opus 4.5 ↓7 0%
assemble_global_geometric_stiffness_CC1_H4_T2
8 models we track have a score here.
- 1
Llama 4 Maverick ↑7 0%
- 2
Gemini 2.5 Pro Jun 2025 ↑1 0%
- 3
Qwen3-Coder-480B-A35B ↑3 0%
- 4 GPT-5 ↑1 0%
- 5 GPT-5 mini ↓1 0%
- 6 Claude Haiku 4.5 ↑1 0%
- 7
Gemini 3 Pro ↓6 0%
- 8 Claude Opus 4.5 ↓6 0%
assemble_global_geometric_stiffness_CC1_H4_T3
8 models we track have a score here.
- 1
Llama 4 Maverick ↑7 0%
- 2
Llama 4 Scout ↑7 0%
- 3
Gemini 2.5 Pro Jun 2025 0%
- 4
Qwen3-Coder-480B-A35B ↑2 0%
- 5 GPT-5 0%
- 6 GPT-5 mini ↓2 0%
- 7 Claude Haiku 4.5 0%
- 8
Gemini 3 Pro ↓7 0%
assemble_global_linear_elastic_stiffness_CC0_H2_T1
9 models we track have a score here.
- 1
Gemini 2.5 Pro Jun 2025 ↑2 100%
- 2
Qwen3-Coder-480B-A35B ↑4 100%
- 3 GPT-5 ↑2 100%
- 4 GPT-5 mini 100%
- 5
Gemini 3 Pro ↓4 100%
- 6 Claude Opus 4.5 ↓4 100%
- 7
Llama 4 Maverick ↑1 0%
- 8
Llama 4 Scout ↑1 0%
- 9 Claude Haiku 4.5 ↓2 0%
assemble_global_linear_elastic_stiffness_CC0_H2_T3
9 models we track have a score here.
- 1
Gemini 2.5 Pro Jun 2025 ↑2 100%
- 2 GPT-5 ↑3 100%
- 3 GPT-5 mini ↑1 100%
- 4
Gemini 3 Pro ↓3 100%
- 5
Llama 4 Maverick ↑3 0%
- 6
Llama 4 Scout ↑3 0%
- 7
Qwen3-Coder-480B-A35B ↓1 0%
- 8 Claude Haiku 4.5 ↓1 0%
- 9 Claude Opus 4.5 ↓7 0%
assemble_global_load_CC0_H0_T0
8 models we track have a score here.
- 1
Llama 4 Maverick ↑7 100%
- 2
Llama 4 Scout ↑7 100%
- 3
Qwen3-Coder-480B-A35B ↑3 100%
- 4 GPT-5 ↑1 100%
- 5 GPT-5 mini ↓1 100%
- 6 Claude Haiku 4.5 ↑1 100%
- 7
Gemini 3 Pro ↓6 100%
- 8 Claude Opus 4.5 ↓6 100%
elastic_critical_load_CC1_H10_T1
9 models we track have a score here.
- 1
Llama 4 Maverick ↑7 0%
- 2
Llama 4 Scout ↑7 0%
- 3
Gemini 2.5 Pro Jun 2025 0%
- 4
Qwen3-Coder-480B-A35B ↑2 0%
- 5 GPT-5 0%
- 6 GPT-5 mini ↓2 0%
- 7 Claude Haiku 4.5 0%
- 8
Gemini 3 Pro ↓7 0%
- 9 Claude Opus 4.5 ↓7 0%
elastic_critical_load_CC1_H10_T2
9 models we track have a score here.
- 1
Llama 4 Maverick ↑7 0%
- 2
Llama 4 Scout ↑7 0%
- 3
Gemini 2.5 Pro Jun 2025 0%
- 4
Qwen3-Coder-480B-A35B ↑2 0%
- 5 GPT-5 0%
- 6 GPT-5 mini ↓2 0%
- 7 Claude Haiku 4.5 0%
- 8
Gemini 3 Pro ↓7 0%
- 9 Claude Opus 4.5 ↓7 0%
elastic_critical_load_CC1_H10_T3
9 models we track have a score here.
- 1
Llama 4 Maverick ↑7 0%
- 2
Llama 4 Scout ↑7 0%
- 3
Gemini 2.5 Pro Jun 2025 0%
- 4
Qwen3-Coder-480B-A35B ↑2 0%
- 5 GPT-5 0%
- 6 GPT-5 mini ↓2 0%
- 7 Claude Haiku 4.5 0%
- 8
Gemini 3 Pro ↓7 0%
- 9 Claude Opus 4.5 ↓7 0%
linear_elastic_CC0_H6_T1
9 models we track have a score here.
- 1
Gemini 2.5 Pro Jun 2025 ↑2 100%
- 2 GPT-5 ↑3 100%
- 3
Gemini 3 Pro ↓2 100%
- 4 Claude Opus 4.5 ↓2 100%
- 5 GPT-5 mini ↓1 50%
- 6 Claude Haiku 4.5 ↑1 50%
- 7
Llama 4 Maverick ↑1 0%
- 8
Llama 4 Scout ↑1 0%
- 9
Qwen3-Coder-480B-A35B ↓3 0%
linear_elastic_CC0_H6_T3
9 models we track have a score here.
- 1 GPT-5 ↑4 100%
- 2 GPT-5 mini ↑2 100%
- 3
Gemini 3 Pro ↓2 100%
- 4 Claude Opus 4.5 ↓2 100%
- 5
Gemini 2.5 Pro Jun 2025 ↓2 50%
- 6 Claude Haiku 4.5 ↑1 50%
- 7
Llama 4 Maverick ↑1 0%
- 8
Llama 4 Scout ↑1 0%
- 9
Qwen3-Coder-480B-A35B ↓3 0%
local_elastic_stiffness_CC0_H0_T0
6 models we track have a score here.
- 1 GPT-5 ↑4 100%
- 2 Claude Opus 4.5 100%
- 3
Gemini 3 Pro ↓2 50%
- 4
Llama 4 Scout ↑5 0%
- 5
Qwen3-Coder-480B-A35B ↑1 0%
- 6 Claude Haiku 4.5 ↑1 0%
local_element_loads_CC0_H2_T1
9 models we track have a score here.
- 1
Gemini 3 Pro 100%
- 2 Claude Opus 4.5 100%
- 3
Llama 4 Maverick ↑5 75%
- 4
Llama 4 Scout ↑5 75%
- 5 GPT-5 mini ↓1 75%
- 6
Gemini 2.5 Pro Jun 2025 ↓3 50%
- 7 GPT-5 ↓2 50%
- 8
Qwen3-Coder-480B-A35B ↓2 0%
- 9 Claude Haiku 4.5 ↓2 0%
local_element_loads_CC0_H2_T3
9 models we track have a score here.
- 1
Gemini 2.5 Pro Jun 2025 ↑2 100%
- 2 GPT-5 mini ↑2 100%
- 3
Gemini 3 Pro ↓2 100%
- 4 Claude Opus 4.5 ↓2 100%
- 5
Llama 4 Maverick ↑3 75%
- 6
Llama 4 Scout ↑3 75%
- 7 GPT-5 ↓2 75%
- 8
Qwen3-Coder-480B-A35B ↓2 0%
- 9 Claude Haiku 4.5 ↓2 0%
local_geometric_stiffness_CC1_H0_T0
8 models we track have a score here.
- 1
Qwen3-Coder-480B-A35B ↑5 50%
- 2 GPT-5 ↑3 50%
- 3 Claude Haiku 4.5 ↑4 50%
- 4
Gemini 3 Pro ↓3 50%
- 5 Claude Opus 4.5 ↓3 50%
- 6
Llama 4 Maverick ↑2 0%
- 7
Llama 4 Scout ↑2 0%
- 8 GPT-5 mini ↓4 0%
partition_DOFs_CC0_H0_T0
9 models we track have a score here.
- 1
Llama 4 Maverick ↑7 100%
- 2
Gemini 2.5 Pro Jun 2025 ↑1 100%
- 3
Qwen3-Coder-480B-A35B ↑3 100%
- 4 GPT-5 ↑1 100%
- 5 GPT-5 mini ↓1 100%
- 6 Claude Haiku 4.5 ↑1 100%
- 7
Gemini 3 Pro ↓6 100%
- 8 Claude Opus 4.5 ↓6 100%
- 9
Llama 4 Scout 0%
solve_eigenvalue_CC1_H1_T1
9 models we track have a score here.
- 1
Llama 4 Maverick ↑7 100%
- 2
Llama 4 Scout ↑7 100%
- 3 GPT-5 ↑2 100%
- 4 GPT-5 mini 100%
- 5
Gemini 3 Pro ↓4 100%
- 6 Claude Opus 4.5 ↓4 100%
- 7
Gemini 2.5 Pro Jun 2025 ↓4 40%
- 8
Qwen3-Coder-480B-A35B ↓2 20%
- 9 Claude Haiku 4.5 ↓2 0%
solve_eigenvalue_CC1_H1_T3
9 models we track have a score here.
- 1
Llama 4 Scout ↑8 100%
- 2 GPT-5 ↑3 100%
- 3 GPT-5 mini ↑1 100%
- 4 Claude Haiku 4.5 ↑3 100%
- 5
Gemini 3 Pro ↓4 100%
- 6 Claude Opus 4.5 ↓4 100%
- 7
Llama 4 Maverick ↑1 80%
- 8
Gemini 2.5 Pro Jun 2025 ↓5 80%
- 9
Qwen3-Coder-480B-A35B ↓3 60%
solve_linear_CC0_H1_T1
9 models we track have a score here.
- 1
Gemini 2.5 Pro Jun 2025 ↑2 100%
- 2
Llama 4 Maverick ↑6 50%
- 3
Llama 4 Scout ↑6 50%
- 4 GPT-5 ↑1 50%
- 5 GPT-5 mini ↓1 50%
- 6
Gemini 3 Pro ↓5 50%
- 7 Claude Opus 4.5 ↓5 50%
- 8
Qwen3-Coder-480B-A35B ↓2 0%
- 9 Claude Haiku 4.5 ↓2 0%
solve_linear_CC0_H1_T3
9 models we track have a score here.
- 1
Gemini 2.5 Pro Jun 2025 ↑2 100%
- 2 GPT-5 ↑3 100%
- 3 GPT-5 mini ↑1 100%
- 4 Claude Haiku 4.5 ↑3 100%
- 5
Gemini 3 Pro ↓4 100%
- 6 Claude Opus 4.5 ↓4 100%
- 7
Qwen3-Coder-480B-A35B ↓1 50%
- 8
Llama 4 Maverick 0%
- 9
Llama 4 Scout 0%
transformation_matrix_CC0_H0_T0
9 models we track have a score here.
- 1
Gemini 3 Pro 100%
- 2
Gemini 2.5 Pro Jun 2025 ↑1 66.7%
- 3 GPT-5 ↑2 66.7%
- 4 Claude Opus 4.5 ↓2 66.7%
- 5
Llama 4 Maverick ↑3 33.3%
- 6
Llama 4 Scout ↑3 33.3%
- 7 GPT-5 mini ↓3 33.3%
- 8 Claude Haiku 4.5 ↓1 33.3%
- 9
Qwen3-Coder-480B-A35B ↓3 0%
2 models have held number one on the official leaderboard since its release in August 2025
Lab claims count from the day the card came out, independent runs from the model’s release.
Source: Official FEM-Bench leaderboard
Gemini 3 Pro on the official leaderboard passed 85% in November 2025
That was 3 months after its release. By the same age, 12% of the benchmarks from 2025 had passed 85%. All the scores we have, one dot each, so a model tested at several settings appears more than once.
Source: Official FEM-Bench 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 Coding
- SWE-bench Verified In use 500 human-validated GitHub issues from 12 Python repos. Agent writes a patch that must pass hidden tests. Claude Opus 4.7 83.5%
- LiveCodeBench v6 In use Competitive programming problems collected after model cutoffs from LeetCode, AtCoder, Codeforces. Scored pass@1 on hidden tests.
DeepSeek-V4-Pro 93.5%
- SWE-bench Multilingual In use 300 GitHub-issue tasks across nine programming languages in SWE-bench format. Patch graded by repository tests. Claude Fable 5.2 96.6%
- DeepSWE In use 113 original long-horizon engineering tasks across 91 repos and five languages, written from scratch. Hand-written verifiers accept alternative solutions. GPT-6 Astra 74.1%
- SciCode In use Scientist-curated research coding problems split into subproblems across natural sciences. Code checked against test cases.
Seed2.1 Pro 59.8%
- Codeforces rating In use Model solves recent Codeforces contest problems. An Elo-style rating is computed from simulated contest performance.
DeepSeek V4.1 Flash 3,471
Data audit
All 8 checks pass. Every benchmark
Catalog record 2 pass
- Pass Task access All tasks are public.
- Pass Confidence Checked against a primary source.
Scores 2 pass 3 notes
- Pass Independent runs Scores from the official leaderboard (10 entries).
- Pass Models scored 9 models we track, plus 1 other entry on its leaderboard.
- Note Newest score The newest came 12 months ago, in November 2025, for Claude Opus 4.5. That’s expected for a saturated test.
- Note Frontier models None of the top 10 has a score, which is fine for a retired test. The strongest tested is Gemini 3 Pro (#56).
- Note Headroom Saturated. The best is 87.9%, past 85% since November 2025.
Leaderboard import 3 pass
- Pass Importer We read the board on github.com: 9 rows for models we track and 1 other entry 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.