Benchmark Almanac

MMCircuitEval

GPT-4V 69.4% leads since July 2025
Not in model cards

Models answer questions about circuit diagrams, specifications, and design decisions. The authors combine text similarity with a language model’s judgment.

OpenReward ranks 33 published model results. ChipOmni separately tested three examples with Qwen.

Released
July 2025
Kind
Electronics
Level
University
Scored
Graded by a model
Answers
Open-ended
Tasks
All public
Measures
Capability
Language
English
In and out
Text, Images → Text
Interaction
Single turn
Refresh
One fixed set
Run by
Anyone
In views
Vision
Strongest model tested
GPT-4 Turbo Apr 2024 #172 on the capability index
Last model added
March 2026
Scores refreshed
November 20th, 2026
33
Entries on OpenReward’s leaderboard, 21 models we track

GPT-4V leads with 69.4% among models released by March 2026 GPT-4V leads with 69.4% among models released by March 2026

Best score per entry from OpenReward’s 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 March 2026, and nobody has tested a newer one. Among the 20 highest-ranked models on the capability index, these have no score yet:

  1. 1 GPT-4V 69.4%
  2. 2 GPT-4o Mar 2025 68%
  3. 3 GPT-4 Turbo Apr 2024 67.4%
  4. 4 ChipExpert* 67.1%
  5. 5 GPT-4 Jun 2023 63.7%
  6. 6 LlaMa3.2-Vision-Instruct-90B 58.5%
  7. 7 Reka Flash 55.7%
  8. 8 GPT-3.5 Turbo Jan 2024 54.9%
  9. 9 QWen-2.5-Instruct-7B 53%
  10. 10 QWen-2-Instruct-72B 50.2%
  11. 11 QWen-2-Instruct-7B 48.2%
  12. 12 LlaMa3.2-Vision-Instruct-11B 43.2%
  13. 13 MiniCPM-Llama3-V 2.5 43%
  14. 14 InternVL2-40B 41.7%
  15. 15 DeepSeekMath-7B-Instruct 40.5%
  16. 16 DeepSeek-Math-7B-RL 39.9%
  17. 17 DeepSeek-V2-Lite-Chat 39.7%
  18. 18 InternVL2-8B 38.8%
  19. 19 Reka Edge 36.7%
  20. 20 DeepSeek-LLM-7B-Chat 35.1%
  21. 21 DeepSeek-MoE-16B-Chat 34%
  22. 22 Yi-VL-34B 32.5%
  23. 23 Qwen-VL-Chat 32.2%
  24. 24 MiniCPM-V 21.1%
  25. 25 InternLM-XComposer-VL-7B 19.7%
  26. 26 Yi-VL-6B 17.3%
  27. 27 BLIP-2 FLAN-T5-XL 14.3%
  28. 28 BLIP-2 FLAN-T5-XXL 14%
  29. 29 QWen-2-Instruct-0.5B 13.8%
  30. 30 Kosmos-2 13.6%
  31. 31 MiniCPM-V 2.0 11.5%
  32. 32 InstructBLIP FLAN-T5-XL 10.8%
  33. 33 InstructBLIP FLAN-T5-XXL 10.7%

Source: OpenReward’s leaderboard

#EntryScoreSettingReleasedRun onLinkSource
1 GPT-4V 69.4% Nov 2023 Jul 2025 arxiv.org Official leaderboard
2 GPT-4o Mar 2025 68% Mar 2025 Jul 2025 arxiv.org Official leaderboard
3 GPT-4 Turbo Apr 2024 67.4% Apr 2024 Jul 2025 arxiv.org Official leaderboard
4 ChipExpert* 67.1% Jul 2025 arxiv.org Official leaderboard
5 GPT-4 Jun 2023 63.7% Jun 2023 Jul 2025 arxiv.org Official leaderboard
6 LlaMa3.2-Vision-Instruct-90B 58.5% Jul 2025 arxiv.org Official leaderboard
7 Reka Flash 55.7% Feb 2024 Jul 2025 arxiv.org Official leaderboard
8 GPT-3.5 Turbo Jan 2024 54.9% Jan 2024 Jul 2025 arxiv.org Official leaderboard
9 QWen-2.5-Instruct-7B 53% Jul 2025 arxiv.org Official leaderboard
10 QWen-2-Instruct-72B 50.2% Jul 2025 arxiv.org Official leaderboard
11 QWen-2-Instruct-7B 48.2% Jul 2025 arxiv.org Official leaderboard
12 LlaMa3.2-Vision-Instruct-11B 43.2% Jul 2025 arxiv.org Official leaderboard
13 MiniCPM-Llama3-V 2.5 43% May 2024 Jul 2025 arxiv.org Official leaderboard
14 InternVL2-40B 41.7% Jul 2024 Jul 2025 arxiv.org Official leaderboard
15 DeepSeekMath-7B-Instruct 40.5% Feb 2024 Jul 2025 arxiv.org Official leaderboard
16 DeepSeek-Math-7B-RL 39.9% Jul 2025 arxiv.org Official leaderboard
17 DeepSeek-V2-Lite-Chat 39.7% Jul 2025 arxiv.org Official leaderboard
18 InternVL2-8B 38.8% Jul 2024 Jul 2025 arxiv.org Official leaderboard
19 Reka Edge 36.7% Mar 2026 Jul 2025 arxiv.org Official leaderboard
20 DeepSeek-LLM-7B-Chat 35.1% Jul 2025 arxiv.org Official leaderboard
21 DeepSeek-MoE-16B-Chat 34% Jul 2025 arxiv.org Official leaderboard
22 Yi-VL-34B 32.5% Dec 2023 Jul 2025 arxiv.org Official leaderboard
23 Qwen-VL-Chat 32.2% Aug 2023 Jul 2025 arxiv.org Official leaderboard
24 MiniCPM-V 21.1% Jan 2024 Jul 2025 arxiv.org Official leaderboard
25 InternLM-XComposer-VL-7B 19.7% Jul 2025 arxiv.org Official leaderboard
26 Yi-VL-6B 17.3% Dec 2023 Jul 2025 arxiv.org Official leaderboard
27 BLIP-2 FLAN-T5-XL 14.3% Feb 2023 Jul 2025 arxiv.org Official leaderboard
28 BLIP-2 FLAN-T5-XXL 14% Feb 2023 Jul 2025 arxiv.org Official leaderboard
29 QWen-2-Instruct-0.5B 13.8% Jul 2025 arxiv.org Official leaderboard
30 Kosmos-2 13.6% Jun 2023 Jul 2025 arxiv.org Official leaderboard
31 MiniCPM-V 2.0 11.5% Apr 2024 Jul 2025 arxiv.org Official leaderboard
32 InstructBLIP FLAN-T5-XL 10.8% May 2023 Jul 2025 arxiv.org Official leaderboard
33 InstructBLIP FLAN-T5-XXL 10.7% Jun 2023 Jul 2025 arxiv.org Official leaderboard
#ModelScoreSettingReleasedRun onLinkSource
1 GPT-4V 69.4% Nov 2023 Jul 2025 arxiv.org Official leaderboard
2 GPT-4o Mar 2025 68% Mar 2025 Jul 2025 arxiv.org Official leaderboard
3 GPT-4 Turbo Apr 2024 67.4% Apr 2024 Jul 2025 arxiv.org Official leaderboard
4 GPT-4 Jun 2023 63.7% Jun 2023 Jul 2025 arxiv.org Official leaderboard
5 Reka Flash 55.7% Feb 2024 Jul 2025 arxiv.org Official leaderboard
6 GPT-3.5 Turbo Jan 2024 54.9% Jan 2024 Jul 2025 arxiv.org Official leaderboard
7 MiniCPM-Llama3-V 2.5 43% May 2024 Jul 2025 arxiv.org Official leaderboard
8 InternVL2-40B 41.7% Jul 2024 Jul 2025 arxiv.org Official leaderboard
9 DeepSeekMath-7B-Instruct 40.5% Feb 2024 Jul 2025 arxiv.org Official leaderboard
10 InternVL2-8B 38.8% Jul 2024 Jul 2025 arxiv.org Official leaderboard
11 Reka Edge 36.7% Mar 2026 Jul 2025 arxiv.org Official leaderboard
12 Yi-VL-34B 32.5% Dec 2023 Jul 2025 arxiv.org Official leaderboard
13 Qwen-VL-Chat 32.2% Aug 2023 Jul 2025 arxiv.org Official leaderboard
14 MiniCPM-V 21.1% Jan 2024 Jul 2025 arxiv.org Official leaderboard
15 Yi-VL-6B 17.3% Dec 2023 Jul 2025 arxiv.org Official leaderboard
16 BLIP-2 FLAN-T5-XL 14.3% Feb 2023 Jul 2025 arxiv.org Official leaderboard
17 BLIP-2 FLAN-T5-XXL 14% Feb 2023 Jul 2025 arxiv.org Official leaderboard
18 Kosmos-2 13.6% Jun 2023 Jul 2025 arxiv.org Official leaderboard
19 MiniCPM-V 2.0 11.5% Apr 2024 Jul 2025 arxiv.org Official leaderboard
20 InstructBLIP FLAN-T5-XL 10.8% May 2023 Jul 2025 arxiv.org Official leaderboard
21 InstructBLIP FLAN-T5-XXL 10.7% Jun 2023 Jul 2025 arxiv.org Official leaderboard

Rankings by part

We read 10 parts from OpenReward’s leaderboard, each ranking models on one part of the test, such as Cph, F and T. GPT-4V leads 5 of them, and 2 other models lead the rest. Arrows show how far a model moved from its overall rank among the models we track.

Cph

21 models we track have a score here.

  1. 1 GPT-4o Mar 2025 ↑1 75.9%
  2. 2 GPT-4V ↓1 75.5%
  3. 3 GPT-4 Turbo Apr 2024 74%
  4. 4 GPT-4 Jun 2023 69.5%
  5. 5 Reka Flash 66.5%
  6. 6 GPT-3.5 Turbo Jan 2024 66.4%
  7. 7 MiniCPM-Llama3-V 2.5 53.5%
  8. 8 InternVL2-40B 51.8%
  9. 9 InternVL2-8B ↑1 47.1%
  10. 10 Qwen-VL-Chat ↑3 46.4%

Number one over time

Lab claims count from the day the card came out, independent runs from the model’s release.

OpenAI
OpenReward leaderboard Not tested since Mar 2026
GPT-4V 69.4%

Source: OpenReward’s leaderboard

Scores over time

All the scores we have, one dot each, so a model tested at several settings appears more than once.

Plot with other benchmarks
On OpenReward’s leaderboard Best so far
0% 20% 40% 60% 80% 100% 2023 2024 2025 2026 Near the ceiling (85%) GPT-4V GPT-4 BLIP-2 FLAN-T5-XL 0% 20% 40% 60% 80% 100% 2023 2024 2025 2026 Near the ceiling (85%) GPT-4V BLIP-2 FLAN-T5-XL

Source: OpenReward’s 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

All 112 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 8 months ago, in March 2026, for Reka Edge. 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 (33 entries).
  • Pass Models scored 21 models we track, plus 12 other entries on its leaderboard.
  • Warning Newest score The newest came 8 months ago, in March 2026, for Reka Edge.
  • Note Frontier models None of the top 10 has a score. The strongest tested is GPT-4 Turbo (#172).
  • Pass Headroom The best is 69.4%, under 85%, so it still separates models.
Leaderboard import 3 pass
  • Pass Importer We read the board on OpenReward: 21 rows for models we track and 12 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.

How we check