NotSoTiny 25-12
Models reconstruct a missing Verilog module from surrounding hardware design code. A formal checker compares the result with the original module.
The authors publish scores for 15 model entries across seven Tiny Tapeout collections. The release contains 1,114 module completion tasks.
- Made by
- Barcelona Supercomputing Center
- Released
- December 2025
- 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
- New versions replace it
- Run by
- Anyone
- Strongest model tested
- Kimi K2.5 #76 on the capability index
- Last model added
- April 2026
- Scores refreshed
- November 20th, 2026
Kimi K2.5 leads with 31.6% among models released by April 2026 Kimi K2.5 leads with 31.6% among models released by April 2026
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 April 2026, 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
- 31.6 Kimi K2.5
- 29.5
Gemma 4 31B IT
- 23.4
Qwen3-Coder-480B-A35B
- 20.9 gpt-oss-120b
- 20.7
DeepSeek-R1
- 19.0 Kimi K2
- 19.0 InCoder-32B
- 14.7
Qwen2.5-72B
- 13.1 InCoder-32B-Thinking
- 12.5
Qwen2.5-14B-Instruct-1M
- 11.8
Qwen2.5-Coder-32B-Instruct
- 8.97 Qwen2.5-14B-Instruct
- 4.20 HaVen-CodeQwen
- 3.48 Qwen2.5-7B-Instruct
- 3.31 OriGen
- 1 Kimi K2.5 31.6%
- 2
Gemma 4 31B IT 29.5%
- 3
Qwen3-Coder-480B-A35B 23.4%
- 4 gpt-oss-120b 20.9%
- 5
DeepSeek-R1 May 2025 20.7%
- 6 Kimi K2 Sep 2025 19%
- 7 InCoder-32B 19%
- 8
Qwen2.5-72B 14.7%
- 9 InCoder-32B-Thinking 13.1%
- 10
Qwen2.5-14B-Instruct-1M 12.5%
- 11
Qwen2.5-Coder-32B-Instruct 11.8%
- 12 Qwen2.5-14B-Instruct 8.97%
- 13 HaVen-CodeQwen 4.20%
- 14 Qwen2.5-7B-Instruct 3.48%
- 15 OriGen 3.31%
Source: Official NotSoTiny 25-12 leaderboard
- 1 Kimi K2.5 31.6%
- 2
Gemma 4 31B IT 29.5%
- 3
Qwen3-Coder-480B-A35B 23.4%
- 4 gpt-oss-120b 20.9%
- 5
DeepSeek-R1 May 2025 20.7%
- 6 Kimi K2 Sep 2025 19%
- 7
Qwen2.5-72B 14.7%
- 8
Qwen2.5-14B-Instruct-1M 12.5%
- 9
Qwen2.5-Coder-32B-Instruct 11.8%
Source: Official NotSoTiny 25-12 leaderboard
| # | Entry | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | Kimi K2.5 | 31.6% | Jan 2026 | Official leaderboard | |
| 2 | | 29.5% | Apr 2026 | Official leaderboard | |
| 3 | | 23.4% | Jul 2025 | Official leaderboard | |
| 4 | gpt-oss-120b | 20.9% | Aug 2025 | Official leaderboard | |
| 5 | | 20.7% | May 2025 | Official leaderboard | |
| 6 | Kimi K2 Sep 2025 | 19% | Sep 2025 | Official leaderboard | |
| 7 | InCoder-32B | 19% | Official leaderboard | ||
| 8 | | 14.7% | Sep 2024 | Official leaderboard | |
| 9 | InCoder-32B-Thinking | 13.1% | Official leaderboard | ||
| 10 | | 12.5% | Jan 2025 | Official leaderboard | |
| 11 | | 11.8% | Nov 2024 | Official leaderboard | |
| 12 | Qwen2.5-14B-Instruct | 8.97% | Official leaderboard | ||
| 13 | HaVen-CodeQwen | 4.20% | Official leaderboard | ||
| 14 | Qwen2.5-7B-Instruct | 3.48% | Official leaderboard | ||
| 15 | OriGen | 3.31% | Official leaderboard |
| # | Model | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | Kimi K2.5 | 31.6% | Jan 2026 | Official leaderboard | |
| 2 | | 29.5% | Apr 2026 | Official leaderboard | |
| 3 | | 23.4% | Jul 2025 | Official leaderboard | |
| 4 | gpt-oss-120b | 20.9% | Aug 2025 | Official leaderboard | |
| 5 | | 20.7% | May 2025 | Official leaderboard | |
| 6 | Kimi K2 Sep 2025 | 19% | Sep 2025 | Official leaderboard | |
| 7 | | 14.7% | Sep 2024 | Official leaderboard | |
| 8 | | 12.5% | Jan 2025 | Official leaderboard | |
| 9 | | 11.8% | Nov 2024 | Official leaderboard |
Rankings by part
We read 7 parts from the benchmark’s own leaderboard, each ranking models on one part of the test, such as TT06, TT07 and TT08. Kimi K2.5 leads 4 of them, and Gemma 4 31B IT leads the other 3. Arrows show how far a model moved from its overall rank among the models we track.
TT06
9 models we track have a score here.
- 1
Gemma 4 31B IT ↑1 22.8%
- 2 Kimi K2.5 ↓1 22.6%
- 3
Qwen3-Coder-480B-A35B 18%
- 4 gpt-oss-120b 17.4%
- 5
DeepSeek-R1 May 2025 14.6%
- 6 Kimi K2 Sep 2025 13%
- 7
Qwen2.5-72B 11.7%
- 8
Qwen2.5-14B-Instruct-1M 9.81%
- 9
Qwen2.5-Coder-32B-Instruct 7.78%
TT07
9 models we track have a score here.
- 1 Kimi K2.5 35%
- 2
Gemma 4 31B IT 31.9%
- 3
Qwen3-Coder-480B-A35B 28.4%
- 4
DeepSeek-R1 May 2025 ↑1 20.6%
- 5 gpt-oss-120b ↓1 19.2%
- 6
Qwen2.5-14B-Instruct-1M ↑2 15.7%
- 7 Kimi K2 Sep 2025 ↓1 15.6%
- 8
Qwen2.5-72B ↓1 12.8%
- 9
Qwen2.5-Coder-32B-Instruct 10.2%
TT08
9 models we track have a score here.
- 1 Kimi K2.5 34.1%
- 2
Gemma 4 31B IT 24.9%
- 3
Qwen3-Coder-480B-A35B 22%
- 4 Kimi K2 Sep 2025 ↑2 20.9%
- 5 gpt-oss-120b ↓1 19.8%
- 6
DeepSeek-R1 May 2025 ↓1 19.7%
- 7
Qwen2.5-14B-Instruct-1M ↑1 15.5%
- 8
Qwen2.5-Coder-32B-Instruct ↑1 5.92%
- 9
Qwen2.5-72B ↓2 5.61%
TT09
9 models we track have a score here.
- 1 Kimi K2.5 42.4%
- 2
Gemma 4 31B IT 36.9%
- 3 gpt-oss-120b ↑1 31.8%
- 4
Qwen3-Coder-480B-A35B ↓1 28.3%
- 5
DeepSeek-R1 May 2025 23.8%
- 6
Qwen2.5-72B ↑1 22.4%
- 7 Kimi K2 Sep 2025 ↓1 21.8%
- 8
Qwen2.5-Coder-32B-Instruct ↑1 17.5%
- 9
Qwen2.5-14B-Instruct-1M ↓1 15.9%
TT10 IHP 02
9 models we track have a score here.
- 1
Gemma 4 31B IT ↑1 26.7%
- 2
DeepSeek-R1 May 2025 ↑3 24.7%
- 3 Kimi K2.5 ↓2 24.7%
- 4 gpt-oss-120b 23.3%
- 5 Kimi K2 Sep 2025 ↑1 23.3%
- 6
Qwen2.5-72B ↑1 22.7%
- 7
Qwen3-Coder-480B-A35B ↓4 20%
- 8
Qwen2.5-Coder-32B-Instruct ↑1 13.3%
- 9
Qwen2.5-14B-Instruct-1M ↓1 11.3%
TT10 IHP 25a
9 models we track have a score here.
- 1
Gemma 4 31B IT ↑1 38%
- 2 Kimi K2.5 ↓1 33.7%
- 3
Qwen3-Coder-480B-A35B 28.1%
- 4
DeepSeek-R1 May 2025 ↑1 20.6%
- 5 Kimi K2 Sep 2025 ↑1 19.6%
- 6 gpt-oss-120b ↓2 18.1%
- 7
Qwen2.5-Coder-32B-Instruct ↑2 15.4%
- 8
Qwen2.5-72B ↓1 15.1%
- 9
Qwen2.5-14B-Instruct-1M ↓1 12.1%
TTSky25a
9 models we track have a score here.
- 1 Kimi K2.5 28.5%
- 2
Gemma 4 31B IT 25.7%
- 3
DeepSeek-R1 May 2025 ↑2 21.1%
- 4
Qwen3-Coder-480B-A35B ↓1 19.2%
- 5 Kimi K2 Sep 2025 ↑1 18.8%
- 6 gpt-oss-120b ↓2 16.6%
- 7
Qwen2.5-Coder-32B-Instruct ↑2 12.8%
- 8
Qwen2.5-72B ↓1 12.7%
- 9
Qwen2.5-14B-Instruct-1M ↓1 7.10%
2 models have held number one on the official leaderboard since its release in December 2025
Lab claims count from the day the card came out, independent runs from the model’s release.
Source: Official NotSoTiny 25-12 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 NotSoTiny 25-12 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 7 months ago, in April 2026, for Gemma 4 31B IT. 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 (15 entries).
- Pass Models scored 9 models we track, plus 6 other entries on its leaderboard.
- Warning Newest score The newest came 7 months ago, in April 2026, for Gemma 4 31B IT.
- Note Frontier models None of the top 10 has a score. The strongest tested is Kimi K2.5 (#76).
- Pass Headroom The best is 31.6%, under 85%, so it still separates models.
Leaderboard import 3 pass
- Pass Importer We read the board on huggingface.co: 9 rows for models we track and 6 other entries 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.