RuC-tt07-32k
Models fill missing Verilog statements in Tiny Tapeout designs. A parser checks syntax, and a formal checker tests whether the completed design preserves behavior.
The authors publish scores for eight models on 851 missing code regions from 30 designs. Some models have scores under two prompting methods.
- Made by
- Barcelona Supercomputing Center
- Released
- April 2026
- Kind
- Electronics
- Level
- University
- Scored
- Checked automatically
- Answers
- Code
- Tasks
- All public
- Measures
- Capability
- Language
- English
- Interaction
- Single turn
- Items from
- Real work and use
- Refresh
- One fixed set
- Run by
- Anyone
- Strongest model tested
- gpt-oss-120b #126 on the capability index
- Last model added
- September 2025
- Scores refreshed
- November 20th, 2026
Qwen3-Coder-480B-A35B leads with 76.7% among models released by September 2025 Qwen3-Coder-480B-A35B leads with 76.7% among models released by September 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. Every run is in the next chart. 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 September 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
Qwen3-Coder-480B-A35B FIM 76.7%
- 2
DeepSeek-V3.1-Terminus FIM 74%
- 3 gpt-oss-120b CHAT 69.6%
- 4 Qwen2.5-Coder-14B FIM 60.2%
- 5 Qwen2.5-Coder-14B-Instruct CHAT 58.9%
- 6 SeedCoder 8B FIM 58.6%
- 7
Qwen3-Coder-30B-A3B-Instruct FIM 48.3%
- 8
Seed-Coder-8B-Instruct FIM 36.3%
Source: Official RuC-tt07-32k leaderboard
- 1
Qwen3-Coder-480B-A35B FIM 76.7%
- 2
DeepSeek-V3.1-Terminus FIM 74%
- 3 gpt-oss-120b CHAT 69.6%
- 4 Qwen2.5-Coder-14B FIM 60.2%
- 5 Qwen2.5-Coder-14B-Instruct CHAT 58.9%
- 6
Qwen3-Coder-30B-A3B-Instruct FIM 48.3%
- 7
Seed-Coder-8B-Instruct FIM 36.3%
Source: Official RuC-tt07-32k leaderboard
| # | Entry | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | | 76.7% | FIM | Jul 2025 | Official leaderboard |
| 2 | | 74% | FIM | Sep 2025 | Official leaderboard |
| 3 | gpt-oss-120b | 69.6% | CHAT | Aug 2025 | Official leaderboard |
| 4 | Qwen2.5-Coder-14B | 60.2% | FIM | Sep 2024 | Official leaderboard |
| 5 | Qwen2.5-Coder-14B-Instruct | 58.9% | CHAT | Nov 2024 | Official leaderboard |
| 6 | SeedCoder 8B | 58.6% | FIM | Official leaderboard | |
| 7 | | 48.3% | FIM | Jul 2025 | Official leaderboard |
| 8 | | 36.3% | FIM | May 2025 | Official leaderboard |
| # | Model | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | | 76.7% | FIM | Jul 2025 | Official leaderboard |
| 2 | | 74% | FIM | Sep 2025 | Official leaderboard |
| 3 | gpt-oss-120b | 69.6% | CHAT | Aug 2025 | Official leaderboard |
| 4 | Qwen2.5-Coder-14B | 60.2% | FIM | Sep 2024 | Official leaderboard |
| 5 | Qwen2.5-Coder-14B-Instruct | 58.9% | CHAT | Nov 2024 | Official leaderboard |
| 6 | | 48.3% | FIM | Jul 2025 | Official leaderboard |
| 7 | | 36.3% | FIM | May 2025 | Official leaderboard |
Rankings by part
We read 18 parts from the benchmark’s own leaderboard in 2 groups: stx and eqv. Each ranks models on one part of the test. Qwen3-Coder-480B-A35B leads 13 of them, and DeepSeek-V3.1-Terminus leads the other 5. Arrows show how far a model moved from its overall rank among the models we track.
INST
7 models we track have a score here.
- 1
Qwen3-Coder-480B-A35B 98%
- 2
DeepSeek-V3.1-Terminus 98%
- 3 Qwen2.5-Coder-14B ↑1 93%
- 4 gpt-oss-120b ↓1 92%
- 5
Qwen3-Coder-30B-A3B-Instruct ↑1 86%
- 6 Qwen2.5-Coder-14B-Instruct ↓1 84%
- 7
Seed-Coder-8B-Instruct 54%
CONT
7 models we track have a score here.
- 1
Qwen3-Coder-480B-A35B 100%
- 2
DeepSeek-V3.1-Terminus 97%
- 3 Qwen2.5-Coder-14B-Instruct ↑2 94%
- 4 gpt-oss-120b ↓1 93%
- 5
Qwen3-Coder-30B-A3B-Instruct ↑1 92%
- 6 Qwen2.5-Coder-14B ↓2 88%
- 7
Seed-Coder-8B-Instruct 64%
BLK
7 models we track have a score here.
- 1
Qwen3-Coder-480B-A35B 100%
- 2
DeepSeek-V3.1-Terminus 96%
- 3
Qwen3-Coder-30B-A3B-Instruct ↑3 91%
- 4 gpt-oss-120b ↓1 90%
- 5 Qwen2.5-Coder-14B ↓1 89%
- 6 Qwen2.5-Coder-14B-Instruct ↓1 87%
- 7
Seed-Coder-8B-Instruct 39%
NBLK
7 models we track have a score here.
- 1
Qwen3-Coder-480B-A35B 100%
- 2
DeepSeek-V3.1-Terminus 98%
- 3 Qwen2.5-Coder-14B ↑1 96%
- 4 Qwen2.5-Coder-14B-Instruct ↑1 89%
- 5 gpt-oss-120b ↓2 89%
- 6
Qwen3-Coder-30B-A3B-Instruct 85%
- 7
Seed-Coder-8B-Instruct 53%
CASE
7 models we track have a score here.
- 1
DeepSeek-V3.1-Terminus ↑1 98%
- 2
Qwen3-Coder-480B-A35B ↓1 94.1%
- 3 Qwen2.5-Coder-14B ↑1 90.2%
- 4 Qwen2.5-Coder-14B-Instruct ↑1 82.3%
- 5 gpt-oss-120b ↓2 76.5%
- 6
Qwen3-Coder-30B-A3B-Instruct 72.5%
- 7
Seed-Coder-8B-Instruct 47.1%
COND
7 models we track have a score here.
- 1
Qwen3-Coder-480B-A35B 98%
- 2 Qwen2.5-Coder-14B ↑2 97%
- 3
DeepSeek-V3.1-Terminus ↓1 96%
- 4 Qwen2.5-Coder-14B-Instruct ↑1 76%
- 5
Qwen3-Coder-30B-A3B-Instruct ↑1 75%
- 6 gpt-oss-120b ↓3 71%
- 7
Seed-Coder-8B-Instruct 50%
ALWS
7 models we track have a score here.
- 1
Qwen3-Coder-480B-A35B 96%
- 2
DeepSeek-V3.1-Terminus 96%
- 3 Qwen2.5-Coder-14B ↑1 89%
- 4
Qwen3-Coder-30B-A3B-Instruct ↑2 87%
- 5 gpt-oss-120b ↓2 85%
- 6 Qwen2.5-Coder-14B-Instruct ↓1 81%
- 7
Seed-Coder-8B-Instruct 62%
PARAM
7 models we track have a score here.
- 1
Qwen3-Coder-480B-A35B 100%
- 2
DeepSeek-V3.1-Terminus 95%
- 3
Qwen3-Coder-30B-A3B-Instruct ↑3 88%
- 4 Qwen2.5-Coder-14B 85%
- 5 gpt-oss-120b ↓2 82%
- 6 Qwen2.5-Coder-14B-Instruct ↓1 74%
- 7
Seed-Coder-8B-Instruct 44%
PORT
7 models we track have a score here.
- 1
Qwen3-Coder-480B-A35B 98%
- 2
DeepSeek-V3.1-Terminus 93%
- 3 gpt-oss-120b 88%
- 4 Qwen2.5-Coder-14B 84%
- 5
Qwen3-Coder-30B-A3B-Instruct ↑1 84%
- 6 Qwen2.5-Coder-14B-Instruct ↓1 59%
- 7
Seed-Coder-8B-Instruct 45%
INST
7 models we track have a score here.
- 1
DeepSeek-V3.1-Terminus ↑1 74%
- 2
Qwen3-Coder-480B-A35B ↓1 70%
- 3 gpt-oss-120b 70%
- 4 Qwen2.5-Coder-14B 69%
- 5
Qwen3-Coder-30B-A3B-Instruct ↑1 66.7%
- 6 Qwen2.5-Coder-14B-Instruct ↓1 61%
- 7
Seed-Coder-8B-Instruct 45%
CONT
7 models we track have a score here.
- 1
Qwen3-Coder-480B-A35B 78%
- 2 gpt-oss-120b ↑1 76%
- 3
DeepSeek-V3.1-Terminus ↓1 76%
- 4
Qwen3-Coder-30B-A3B-Instruct ↑2 62%
- 5 Qwen2.5-Coder-14B-Instruct 60%
- 6 Qwen2.5-Coder-14B ↓2 55%
- 7
Seed-Coder-8B-Instruct 36%
BLK
7 models we track have a score here.
- 1
Qwen3-Coder-480B-A35B 88%
- 2
DeepSeek-V3.1-Terminus 85%
- 3 gpt-oss-120b 81%
- 4 Qwen2.5-Coder-14B-Instruct ↑1 73%
- 5 Qwen2.5-Coder-14B ↓1 65%
- 6
Qwen3-Coder-30B-A3B-Instruct 49%
- 7
Seed-Coder-8B-Instruct 32%
NBLK
7 models we track have a score here.
- 1
Qwen3-Coder-480B-A35B 87%
- 2
DeepSeek-V3.1-Terminus 86%
- 3 gpt-oss-120b 79%
- 4 Qwen2.5-Coder-14B-Instruct ↑1 74%
- 5 Qwen2.5-Coder-14B ↓1 69%
- 6
Qwen3-Coder-30B-A3B-Instruct 62%
- 7
Seed-Coder-8B-Instruct 43%
CASE
7 models we track have a score here.
- 1
DeepSeek-V3.1-Terminus ↑1 66.7%
- 2
Qwen3-Coder-480B-A35B ↓1 56.9%
- 3 gpt-oss-120b 49%
- 4 Qwen2.5-Coder-14B-Instruct ↑1 43.1%
- 5
Qwen3-Coder-30B-A3B-Instruct ↑1 37.2%
- 6 Qwen2.5-Coder-14B ↓2 33.3%
- 7
Seed-Coder-8B-Instruct 33.3%
COND
7 models we track have a score here.
- 1
DeepSeek-V3.1-Terminus ↑1 67%
- 2
Qwen3-Coder-480B-A35B ↓1 63%
- 3 gpt-oss-120b 54%
- 4 Qwen2.5-Coder-14B-Instruct ↑1 53%
- 5
Qwen3-Coder-30B-A3B-Instruct ↑1 52.9%
- 6 Qwen2.5-Coder-14B ↓2 44%
- 7
Seed-Coder-8B-Instruct 33%
ALWS
7 models we track have a score here.
- 1
DeepSeek-V3.1-Terminus ↑1 60%
- 2
Qwen3-Coder-480B-A35B ↓1 57%
- 3 gpt-oss-120b 53%
- 4 Qwen2.5-Coder-14B-Instruct ↑1 47%
- 5
Qwen3-Coder-30B-A3B-Instruct ↑1 39%
- 6 Qwen2.5-Coder-14B ↓2 31%
- 7
Seed-Coder-8B-Instruct 31%
PARAM
7 models we track have a score here.
- 1
Qwen3-Coder-480B-A35B 92%
- 2
DeepSeek-V3.1-Terminus 88%
- 3 Qwen2.5-Coder-14B ↑1 81%
- 4 gpt-oss-120b ↓1 74%
- 5 Qwen2.5-Coder-14B-Instruct 67%
- 6
Qwen3-Coder-30B-A3B-Instruct 63%
- 7
Seed-Coder-8B-Instruct 43%
PORT
7 models we track have a score here.
- 1
Qwen3-Coder-480B-A35B 89%
- 2
DeepSeek-V3.1-Terminus 85%
- 3 Qwen2.5-Coder-14B ↑1 81%
- 4 gpt-oss-120b ↓1 80%
- 5 Qwen2.5-Coder-14B-Instruct 54%
- 6
Qwen3-Coder-30B-A3B-Instruct 50%
- 7
Seed-Coder-8B-Instruct 41%
Scores at each effort level
Models the leaderboard ran at two or more effort levels. We don’t have enough cost data here to say what the extra effort costs.
Source: Official RuC-tt07-32k leaderboard
| Model | Effort | Harness | Score | Cost per task | Output tokens |
|---|---|---|---|---|---|
| FIM | 76.7% | ||||
| FIM | 74% | ||||
| CHAT | 73.9% | ||||
| CHAT | 71% | ||||
| gpt-oss-120b | CHAT | 69.6% | |||
| Qwen2.5-Coder-14B | FIM | 60.2% | |||
| Qwen2.5-Coder-14B-Instruct | CHAT | 58.9% | |||
| Qwen2.5-Coder-14B-Instruct | FIM | 49.2% | |||
| FIM | 48.3% | ||||
| CHAT | 46.5% | ||||
| FIM | 36.3% | ||||
| CHAT | 23.6% |
Scores at each effort level
Models the leaderboard ran at two or more effort levels. We don’t have enough cost data here to say what the extra effort costs.
Source: Official RuC-tt07-32k leaderboard
| Model | Effort | Harness | Score | Cost per task | Output tokens |
|---|---|---|---|---|---|
| FIM | 76.7% | ||||
| FIM | 74% | ||||
| CHAT | 73.9% | ||||
| CHAT | 71% | ||||
| gpt-oss-120b | CHAT | 69.6% | |||
| Qwen2.5-Coder-14B | FIM | 60.2% | |||
| Qwen2.5-Coder-14B-Instruct | CHAT | 58.9% | |||
| Qwen2.5-Coder-14B-Instruct | FIM | 49.2% | |||
| FIM | 48.3% | ||||
| CHAT | 46.5% | ||||
| FIM | 36.3% | ||||
| CHAT | 23.6% |
Number one over time
Lab claims count from the day the card came out, independent runs from the model’s release.
Source: Official RuC-tt07-32k leaderboard
Scores over time
All the scores we have, one dot each, so a model tested at several settings appears more than once.
Source: Official RuC-tt07-32k 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 Engineering and CAD
- EEBench In use 13 electrical-engineering circuit and PCB design tasks graded by deterministic build checks, simulation and BOM constraints. No LLM judge.
Gemini 4 Carbon 83.9%
- 3DCodeBench In use Agentic procedural 3D modeling via code: author engine-ready 3D assets through software APIs. Graded for executability and shape fidelity.
LongCat-2.5 55.4%
- CADGenBench Only in system cards Agents construct CAD models from design prompts. The generation split is graded for executability and geometric correctness. Claude Fable 5 45.1%
- PartBench Only in system cards Parametric mechanical part design and CAD-style generation, graded for geometric and manufacturing fidelity. Grok 4.6 58.5%
- Razavi-Bench Not in model cards Models explain what happens in analog circuits and why. A language model awards up to four points per answer against a reference and a rubric.
Gemini 4 Carbon 98.3%
- CircuitRubric Not in model cards Models write analog circuit netlists from design requests. A program compares transistor connections and relative device sizes with accepted reference circuits. Claude Opus 4.8 96%
Data audit
9 of 10 checks pass, 1 warning. Look first at the Newest score check. The newest came 14 months ago, in September 2025, for DeepSeek-V3.1-Terminus. Every benchmark
Catalog record 2 pass
- Pass Task access All tasks are public.
- Pass Confidence Checked against a primary source.
Scores 1 warning 3 pass 1 note
- Pass Independent runs Scores from the official leaderboard (8 entries).
- Pass Models scored 7 models we track, plus 1 other entry on its leaderboard.
- Warning Newest score The newest came 14 months ago, in September 2025, for DeepSeek-V3.1-Terminus.
- Note Frontier models None of the top 10 has a score. The strongest tested is gpt-oss-120b (#126).
- Pass Headroom The best is 76.7%, under 85%, so it still separates models.
Leaderboard import 3 pass
- Pass Importer We read the board on huggingface.co: 12 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.