MultiMedia-TerminalBench
Agents inspect and edit media in persistent terminal workspaces. The 105 tasks require verifiable output files across five areas of professional multimedia work.
In May 2026, the best Gemini 3.1 Pro setup reached 37.1% binary success and 46.9% partial credit. GPT-5.2 with Codex CLI solved 16.2%.
- Made by
- Chiyeong Heo and colleagues
- Released
- May 2026
- Kind
- Video understanding
- Level
- Expert
- Scored
- Checked automatically
- Answers
- Agent task
- Tasks
- All public
- Measures
- Capability
- Subject
- Arts, Media and communications
- Language
- English
- Interaction
- Agentic
- Environment
- Terminal
- Refresh
- One fixed set
- Run by
- Anyone
- Occupations
- Arts, design, entertainment, sports and media
- In views
- Agents, Vision, Audio and speech, Image, video and 3D generation
- Strongest model tested
-
Gemini 3.1 Pro #40 on the capability index
- Last model added
- February 2026
- Scores refreshed
- November 20th, 2026
Terminus-MM + Gemini 3.1 Pro tops the leaderboard with 37.1% 2 models share the top score of 16.2% among models released by February 2026 Gemini 3.1 Pro leads with 12.4% in Terminus 2
Best score per entry from the benchmark's paper, read on November 20th, 2026. A paper’s table doesn’t add newer models. Where a model ran on its own with several settings, we keep its best. Every run is in the next chart. Agents other companies built around a model get a pale shade of that model's lab color. Gray marks blends and entries not in our model list, such as fine-tunes and older models. None of these count toward leaders. The best model on its own is GPT-5.2, at 16.2%. Agents built around a model, such as Terminus-MM + Gemini 3.1 Pro, are their own entries. We don't count them as the model's score. This view shows only models we track, each at its best. This view keeps only runs in Terminus 2, a harness no model’s maker built. A maker’s own agent, such as Claude Code with a Claude model, is left out.
The newest model with a score here came out in February 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
- 37.1 Terminus-MM + Gemini 3.1 Pro
- 33.3 Terminus-IA + Gemini 3.1 Pro
- 33.3 Terminus-IV + Gemini 3.1 Pro
- 28.6 Terminus-V + Gemini 3.1 Pro
- 27.6 Terminus-AV + Gemini 3.1 Pro
- 25.7 Terminus-A + Gemini 3.1 Pro
- 22.9 Terminus-V + Gemini 2.5 Flash
- 22.9 Terminus-MM + Gemini 2.5 Flash
- 21.9 Terminus-AV + Gemini 2.5 Flash
- 16.2 Terminus-IV + Gemini 2.5 Flash
- 16.2 GPT-5.2
- 16.2 Claude Sonnet 4.6
- 13.3 Terminus-IA + Gemini 2.5 Flash
- 12.4
Gemini 3.1 Pro
- 11.4 Terminus-KIRA + GPT-5.2
- 10.5 Terminus-A + Gemini 2.5 Flash
- 10.5 Terminus-KIRA + Gemini 3.1 Pro
- 10.5
Qwen3.5-122B-A10B
- 9.50 Terminus-KIRA + Qwen3.5-122B-A10B
- 6.70
Gemini 2.5 Flash
- 6.70 Terminus-KIRA + Gemini 2.5 Flash
- 1 Terminus-MM + Gemini 3.1 Pro T+I+A+V 37.1%
- 2 Terminus-IA + Gemini 3.1 Pro T+I+A 33.3%
- 3 Terminus-IV + Gemini 3.1 Pro T+I+V 33.3%
- 4 Terminus-V + Gemini 3.1 Pro T+V 28.6%
- 5 Terminus-AV + Gemini 3.1 Pro T+A+V 27.6%
- 6 Terminus-A + Gemini 3.1 Pro T+A 25.7%
- 7 Terminus-V + Gemini 2.5 Flash Sep 2025, T+V 22.9%
- 8 Terminus-MM + Gemini 2.5 Flash Sep 2025, T+I+A+V 22.9%
- 9 Terminus-AV + Gemini 2.5 Flash Sep 2025, T+A+V 21.9%
- 10 Terminus-IV + Gemini 2.5 Flash Sep 2025, T+I+V 16.2%
- 11 GPT-5.2 T+I 16.2%
- 12 Claude Sonnet 4.6 T+I 16.2%
- 13 Terminus-IA + Gemini 2.5 Flash Sep 2025, T+I+A 13.3%
- 14
Gemini 3.1 Pro T 12.4%
- 15 Terminus-KIRA + GPT-5.2 T+I 11.4%
- 16 Terminus-A + Gemini 2.5 Flash Sep 2025, T+A 10.5%
- 17 Terminus-KIRA + Gemini 3.1 Pro T+I 10.5%
- 18
Qwen3.5-122B-A10B T 10.5%
- 19 Terminus-KIRA + Qwen3.5-122B-A10B T+I 9.50%
- 20
Gemini 2.5 Flash Sep 2025, T 6.70%
- 21 Terminus-KIRA + Gemini 2.5 Flash Sep 2025, T+I 6.70%
Source: The MultiMedia-TerminalBench paper
- 1 GPT-5.2 T+I 16.2%
- 2 Claude Sonnet 4.6 T+I 16.2%
- 3
Gemini 3.1 Pro T 12.4%
- 4
Qwen3.5-122B-A10B T 10.5%
- 5
Gemini 2.5 Flash Sep 2025, T 6.70%
Source: The MultiMedia-TerminalBench paper
- 1
Gemini 3.1 Pro T 12.4%
- 2 GPT-5.2 T 10.5%
- 3
Qwen3.5-122B-A10B T 10.5%
- 4
Gemini 2.5 Flash Sep 2025, T 6.70%
Source: The MultiMedia-TerminalBench paper
| # | Entry | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | Terminus-MM + Gemini 3.1 Pro | 37.1% | T+I+A+V | Feb 2026 | Official leaderboard |
| 2 | Terminus-IA + Gemini 3.1 Pro | 33.3% | T+I+A | Feb 2026 | Official leaderboard |
| 3 | Terminus-IV + Gemini 3.1 Pro | 33.3% | T+I+V | Feb 2026 | Official leaderboard |
| 4 | Terminus-V + Gemini 3.1 Pro | 28.6% | T+V | Feb 2026 | Official leaderboard |
| 5 | Terminus-AV + Gemini 3.1 Pro | 27.6% | T+A+V | Feb 2026 | Official leaderboard |
| 6 | Terminus-A + Gemini 3.1 Pro | 25.7% | T+A | Feb 2026 | Official leaderboard |
| 7 | Terminus-V + Gemini 2.5 Flash Sep 2025 | 22.9% | T+V | Sep 2025 | Official leaderboard |
| 8 | Terminus-MM + Gemini 2.5 Flash Sep 2025 | 22.9% | T+I+A+V | Sep 2025 | Official leaderboard |
| 9 | Terminus-AV + Gemini 2.5 Flash Sep 2025 | 21.9% | T+A+V | Sep 2025 | Official leaderboard |
| 10 | Terminus-IV + Gemini 2.5 Flash Sep 2025 | 16.2% | T+I+V | Sep 2025 | Official leaderboard |
| 11 | GPT-5.2 | 16.2% | T+I, Codex CLI | Dec 2025 | Official leaderboard |
| 12 | Claude Sonnet 4.6 | 16.2% | T+I, Claude Code | Feb 2026 | Official leaderboard |
| 13 | Terminus-IA + Gemini 2.5 Flash Sep 2025 | 13.3% | T+I+A | Sep 2025 | Official leaderboard |
| 14 | | 12.4% | T, Terminus-2 | Feb 2026 | Official leaderboard |
| 15 | Terminus-KIRA + GPT-5.2 | 11.4% | T+I | Dec 2025 | Official leaderboard |
| 16 | Terminus-A + Gemini 2.5 Flash Sep 2025 | 10.5% | T+A | Sep 2025 | Official leaderboard |
| 17 | Terminus-KIRA + Gemini 3.1 Pro | 10.5% | T+I | Feb 2026 | Official leaderboard |
| 18 | | 10.5% | T, Terminus-2 | Feb 2026 | Official leaderboard |
| 19 | Terminus-KIRA + Qwen3.5-122B-A10B | 9.50% | T+I | Feb 2026 | Official leaderboard |
| 20 | | 6.70% | T, Terminus-2 | Sep 2025 | Official leaderboard |
| 21 | Terminus-KIRA + Gemini 2.5 Flash Sep 2025 | 6.70% | T+I | Sep 2025 | Official leaderboard |
| # | Model | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | GPT-5.2 | 16.2% | T+I, Codex CLI | Dec 2025 | Official leaderboard |
| 2 | Claude Sonnet 4.6 | 16.2% | T+I, Claude Code | Feb 2026 | Official leaderboard |
| 3 | | 12.4% | T, Terminus-2 | Feb 2026 | Official leaderboard |
| 4 | | 10.5% | T, Terminus-2 | Feb 2026 | Official leaderboard |
| 5 | | 6.70% | T, Terminus-2 | Sep 2025 | Official leaderboard |
| # | Model | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | | 12.4% | T, Terminus-2 | Feb 2026 | Official leaderboard |
| 2 | GPT-5.2 | 10.5% | T, Terminus-2 | Dec 2025 | Official leaderboard |
| 3 | | 10.5% | T, Terminus-2 | Feb 2026 | Official leaderboard |
| 4 | | 6.70% | T, Terminus-2 | Sep 2025 | Official leaderboard |
Rankings by part
We read 17 parts from the benchmark’s paper in 2 groups: capability and meta-category. Each ranks models on one part of the test. GPT-5.2 leads 8 of them, and 3 other models lead the rest. Arrows show how far a model moved from its overall rank among the models we track.
Audio-Visual Alignment
5 models we track have a score here.
- 1 Claude Sonnet 4.6 ↑1 12.7%
- 2 GPT-5.2 ↓1 10.9%
- 3
Gemini 2.5 Flash Sep 2025 ↑2 9.10%
- 4
Gemini 3.1 Pro ↓1 9.10%
- 5
Qwen3.5-122B-A10B ↓1 3.60%
Cross-File Comparison
5 models we track have a score here.
- 1 Claude Sonnet 4.6 ↑1 28.6%
- 2 GPT-5.2 ↓1 23.8%
- 3
Gemini 3.1 Pro 19.1%
- 4
Qwen3.5-122B-A10B 19.1%
- 5
Gemini 2.5 Flash Sep 2025 14.3%
Music Understanding
5 models we track have a score here.
- 1 GPT-5.2 9.10%
- 2 Claude Sonnet 4.6 9.10%
- 3
Gemini 2.5 Flash Sep 2025 ↑2 0%
- 4
Gemini 3.1 Pro ↓1 0%
- 5
Qwen3.5-122B-A10B ↓1 0%
Non-Speech Audio
5 models we track have a score here.
- 1 Claude Sonnet 4.6 ↑1 26.9%
- 2
Gemini 3.1 Pro ↑1 23.1%
- 3 GPT-5.2 ↓2 19.2%
- 4
Qwen3.5-122B-A10B 11.5%
- 5
Gemini 2.5 Flash Sep 2025 3.80%
On-Screen Text
5 models we track have a score here.
- 1 GPT-5.2 16.7%
- 2 Claude Sonnet 4.6 16.7%
- 3
Gemini 3.1 Pro 11.9%
- 4
Qwen3.5-122B-A10B 9.50%
- 5
Gemini 2.5 Flash Sep 2025 0%
Reference Resolution
5 models we track have a score here.
- 1 GPT-5.2 14.3%
- 2 Claude Sonnet 4.6 14.3%
- 3
Gemini 2.5 Flash Sep 2025 ↑2 0%
- 4
Gemini 3.1 Pro ↓1 0%
- 5
Qwen3.5-122B-A10B ↓1 0%
Spatial Reasoning
5 models we track have a score here.
- 1 Claude Sonnet 4.6 ↑1 41.7%
- 2 GPT-5.2 ↓1 25%
- 3
Gemini 3.1 Pro 25%
- 4
Qwen3.5-122B-A10B 16.7%
- 5
Gemini 2.5 Flash Sep 2025 8.30%
Speaker/Voice Identity
5 models we track have a score here.
- 1 GPT-5.2 15.4%
- 2 Claude Sonnet 4.6 15.4%
- 3
Gemini 3.1 Pro 15.4%
- 4
Gemini 2.5 Flash Sep 2025 ↑1 7.70%
- 5
Qwen3.5-122B-A10B ↓1 7.70%
Speech Prosody
5 models we track have a score here.
- 1
Gemini 2.5 Flash Sep 2025 ↑4 15.4%
- 2
Qwen3.5-122B-A10B ↑2 15.4%
- 3 GPT-5.2 ↓2 7.70%
- 4 Claude Sonnet 4.6 ↓2 7.70%
- 5
Gemini 3.1 Pro ↓2 7.70%
Speech Understanding
5 models we track have a score here.
- 1 GPT-5.2 12.3%
- 2 Claude Sonnet 4.6 8.80%
- 3
Qwen3.5-122B-A10B ↑1 8.80%
- 4
Gemini 2.5 Flash Sep 2025 ↑1 7.00%
- 5
Gemini 3.1 Pro ↓2 5.30%
Temporal Localization
5 models we track have a score here.
- 1 GPT-5.2 15.9%
- 2
Gemini 3.1 Pro ↑1 15.9%
- 3
Qwen3.5-122B-A10B ↑1 11.4%
- 4 Claude Sonnet 4.6 ↓2 9.10%
- 5
Gemini 2.5 Flash Sep 2025 6.80%
Visual Perception
5 models we track have a score here.
- 1 Claude Sonnet 4.6 ↑1 23.5%
- 2 GPT-5.2 ↓1 22.1%
- 3
Gemini 3.1 Pro 16.2%
- 4
Qwen3.5-122B-A10B 8.80%
- 5
Gemini 2.5 Flash Sep 2025 7.30%
Media Production
5 models we track have a score here.
- 1 GPT-5.2 22.5%
- 2 Claude Sonnet 4.6 15%
- 3
Gemini 3.1 Pro 12.5%
- 4
Gemini 2.5 Flash Sep 2025 ↑1 7.50%
- 5
Qwen3.5-122B-A10B ↓1 7.50%
Performance & Coaching
5 models we track have a score here.
- 1
Qwen3.5-122B-A10B ↑3 13%
- 2
Gemini 2.5 Flash Sep 2025 ↑3 8.70%
- 3
Gemini 3.1 Pro 8.70%
- 4 GPT-5.2 ↓3 4.30%
- 5 Claude Sonnet 4.6 ↓3 4.30%
Enterprise & Compliance
5 models we track have a score here.
- 1 Claude Sonnet 4.6 ↑1 21.7%
- 2 GPT-5.2 ↓1 17.4%
- 3
Gemini 3.1 Pro 13%
- 4
Qwen3.5-122B-A10B 13%
- 5
Gemini 2.5 Flash Sep 2025 0%
Personal & Education
5 models we track have a score here.
- 1 Claude Sonnet 4.6 ↑1 30.8%
- 2 GPT-5.2 ↓1 15.4%
- 3
Gemini 3.1 Pro 15.4%
- 4
Qwen3.5-122B-A10B 15.4%
- 5
Gemini 2.5 Flash Sep 2025 7.70%
Operations & Research
5 models we track have a score here.
- 1 GPT-5.2 33.3%
- 2
Gemini 2.5 Flash Sep 2025 ↑3 16.7%
- 3 Claude Sonnet 4.6 ↓1 16.7%
- 4
Gemini 3.1 Pro ↓1 16.7%
- 5
Qwen3.5-122B-A10B ↓1 0%
The best score, 37.1%, costs $1.23 a task
Runs on the cost frontier score higher than any run that's cheaper. The costs are the leaderboard's own estimates, so compare them only within this benchmark.
Source: The MultiMedia-TerminalBench paper
| Model | Effort | Harness | Score | Cost per task | Output tokens |
|---|---|---|---|---|---|
| Terminus-MM + Gemini 3.1 Pro | T+I+A+V | Terminus-MM | 37.1% | $1.23 | |
| Terminus-IA + Gemini 3.1 Pro | T+I+A | Terminus-IA | 33.3% | $1.74 | |
| Terminus-IV + Gemini 3.1 Pro | T+I+V | Terminus-IV | 33.3% | $1.28 | |
| Terminus-MM + Gemini 3.1 Pro | Without modality masking | Terminus-MM | 32.4% | ||
| Terminus-V + Gemini 3.1 Pro | T+V | Terminus-V | 28.6% | $1.12 | |
| Terminus-AV + Gemini 3.1 Pro | T+A+V | Terminus-AV | 27.6% | $1.03 | |
| Terminus-A + Gemini 3.1 Pro | T+A | Terminus-A | 25.7% | $1.99 | |
| Terminus-V + Gemini 2.5 Flash | T+V | Terminus-V | 22.9% | $0.14 | |
| Terminus-MM + Gemini 2.5 Flash | T+I+A+V | Terminus-MM | 22.9% | $0.10 | |
| Terminus-AV + Gemini 2.5 Flash | T+A+V | Terminus-AV | 21.9% | $0.14 | |
| Terminus-MM + Gemini 2.5 Flash | Without modality masking | Terminus-MM | 17.1% | ||
| GPT-5.2 | T+I | Codex CLI | 16.2% | $7.12 | |
| Claude Sonnet 4.6 | T+I | Claude Code | 16.2% | $1.74 | |
| Terminus-IV + Gemini 2.5 Flash | T+I+V | Terminus-IV | 16.2% | $0.18 | |
| Terminus-IA + Gemini 2.5 Flash | T+I+A | Terminus-IA | 13.3% | $0.23 | |
| T | Terminus-2 | 12.4% | $0.77 | ||
| Terminus-KIRA + GPT-5.2 | T+I | Terminus-KIRA | 11.4% | $1.67 | |
| GPT-5.2 | T | Terminus-2 | 10.5% | $0.82 | |
| T | Terminus-2 | 10.5% | $0.10 | ||
| Terminus-A + Gemini 2.5 Flash | T+A | Terminus-A | 10.5% | $0.13 | |
| Terminus-KIRA + Gemini 3.1 Pro | T+I | Terminus-KIRA | 10.5% | $2.06 | |
| Terminus-KIRA + Qwen3.5-122B-A10B | T+I | Terminus-KIRA | 9.50% | $0.23 | |
| T | Terminus-2 | 6.70% | $0.12 | ||
| Terminus-KIRA + Gemini 2.5 Flash | T+I | Terminus-KIRA | 6.70% | $0.23 |
Number one over time
Lab claims count from the day the card came out, independent runs from the model’s release.
Source: The MultiMedia-TerminalBench paper
Scores over time
All the scores we have, one dot each, so a model tested at several settings appears more than once.
Source: Official MultiMedia-TerminalBench leaderboard
Source: Official MultiMedia-TerminalBench 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 Audio and video
- Video-MMMU In use 300 expert-level lecture videos with questions testing perception, comprehension and adaptation of knowledge. Accuracy.
Qwen 3.8 Max 88.7%
- LVBench In use Hours-long video understanding with multiple-choice questions on reasoning, grounding and summarization.
Gemini 4 Argon 91.7%
- MMVU In use 3,000 expert-annotated questions on videos across 27 disciplines. Multiple-choice and open-ended accuracy.
Qwen 3.8 Max 82.4%
- MLVU In use Multi-task long video understanding (minutes to hours) with multiple-choice and generation tasks. Accuracy.
Qwen 3.8 Max 90.8%
- FLEURS In use Models transcribe read speech in 102 languages built from FLoRes-101 sentences. Model cards usually report word error rate, so lower scores are better.
Phi-4-multimodal 4.00%
- VoiceBench In use Spoken instructions with speaker, environment and content variation. Voice assistant responses scored by GPT judge or rules.
Qwen3.5-Omni-Plus 93.1%
Data audit
9 of 10 checks pass, 1 warning. Look first at the Newest score check. The newest came 9 months ago, in February 2026, for Qwen3.5-122B-A10B. 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 benchmark’s paper (21 entries).
- Pass Models scored 5 models we track, plus 16 other entries on its leaderboard.
- Warning Newest score The newest came 9 months ago, in February 2026, for Qwen3.5-122B-A10B.
- Note Frontier models None of the top 10 has a score. The strongest tested is Gemini 3.1 Pro (#40).
- Pass Headroom The best is 16.2%, under 85%, so it still separates models.
Leaderboard import 3 pass
- Pass Importer We read the board on arxiv.org: 24 rows for models we track 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.