# Terminal-Bench 4.1

> The current Terminal-Bench release keeps the 66 professional computer-work tasks of 4.0 and regrades every saved trial with verifiers hardened against reward hacking and overly strict checks.

Canonical page: https://benchlm.ai/benchmarks/terminal-bench

- Category: [Agentic](/agentic)
- Last updated: November 10, 2026 snapshot

## About Terminal-Bench 4.1

- Year: 2026
- Tasks: 66 professional computer-work tasks
- Format: Task completion rate across 5 trials per task
- Difficulty: Frontier autonomous knowledge work
- Paper: [Terminal-Bench 4.1](https://www.tbench.ai/news/terminal-bench-4-1)

Version 4.1 makes no agent-facing changes. It re-scores the saved 4.0 artifacts with updated verifiers: agent code now runs as an untrusted user in a separate process, and several over-strict checks accept valid alternative solutions. 162 of 11,550 trials changed score, and every one of the 35 leaderboard entries stayed within its 4.0 95% confidence interval. BenchLM keeps 4.1 on its own key so the 4.0 table remains a fixed record.

Terminal-Bench 4.1 is currently displayed on BenchLM for reference, but it is excluded from the weighted scoring formula.

## Leaderboard (35 models)

| Rank | Model | Configuration | Creator | Score |
|------|-------|---------------|---------|-------|
| 1 | [Claude Opus 5.5](/models/claude-opus-5-5) | Claude Code · max reasoning | Anthropic | 65.76% |
| 2 | [Claude Sonnet 5.5](/models/claude-sonnet-5-5) | Claude Code · max reasoning | Anthropic | 63.94% |
| 3 | [GPT-6.1 Sol](/models/gpt-6-1-sol) | Codex · max reasoning | OpenAI | 60.91% |
| 4 | [GPT-6 Astra](/models/gpt-6-astra) | Codex · max reasoning | OpenAI | 60.30% |
| 5 | [GPT-6 Astra](/models/gpt-6-astra) | Codex · high reasoning | OpenAI | 60.00% |
| 6 | [GPT-6 Astra](/models/gpt-6-astra) | Codex · xhigh reasoning | OpenAI | 59.70% |
| 7 | [Claude Fable 5.1](/models/claude-fable-5-1) | Claude Code · xhigh reasoning | Anthropic | 57.58% |
| 8 | [Claude Fable 5.1](/models/claude-fable-5-1) | Claude Code · max reasoning | Anthropic | 57.27% |
| 9 | [GPT-6 Astra](/models/gpt-6-astra) | Codex · medium reasoning | OpenAI | 55.15% |
| 10 | [Claude Fable 5.1](/models/claude-fable-5-1) | Claude Code · high reasoning | Anthropic | 54.85% |
| 11 | [Claude Fable 5.1](/models/claude-fable-5-1) | Claude Code · medium reasoning | Anthropic | 53.94% |
| 12 | [Claude Opus 5](/models/claude-opus-5) | Claude Code · xhigh reasoning | Anthropic | 53.33% |
| 13 | [Claude Opus 5](/models/claude-opus-5) | Claude Code · max reasoning | Anthropic | 52.12% |
| 14 | [GPT-6 Astra](/models/gpt-6-astra) | Codex · low reasoning | OpenAI | 50.91% |
| 15 | [GPT-6 Sol](/models/gpt-6-sol) | Codex · max reasoning | OpenAI | 50.30% |
| 16 | [Claude Opus 5](/models/claude-opus-5) | Claude Code · high reasoning | Anthropic | 49.70% |
| 17 | [Claude Opus 5](/models/claude-opus-5) | Claude Code · medium reasoning | Anthropic | 45.15% |
| 18 | [Claude Fable 5](/models/claude-fable) | Claude Code · max reasoning | Anthropic | 44.85% |
| 19 | [Claude Fable 5.1](/models/claude-fable-5-1) | Claude Code · low reasoning | Anthropic | 43.33% |
| 20 | [GLM-5.3](/models/glm-5-3) | Claude Code · max reasoning | Z.AI | 41.21% |
| 21 | [Grok 4.7](/models/grok-4-7) | Grok Build · xhigh reasoning | xAI | 38.79% |
| 22 | [GPT-5.6 Sol](/models/gpt-5-6-sol) | Codex · max reasoning | OpenAI | 38.48% |
| 23 | [GLM-5.3-Flash](/models/glm-5-3-flash) | Claude Code · none reasoning | Z.AI | 35.45% |
| 24 | [Claude Opus 5](/models/claude-opus-5) | Claude Code · low reasoning | Anthropic | 34.85% |
| 25 | [Qwen3.8-Max-0902](https://www.tbench.ai/leaderboard/terminal-bench/4.1) | Claude Code · max reasoning | Alibaba | 26.97% |
| 26 | [Claude Opus 4.8](/models/claude-opus-4-8) | Claude Code · max reasoning | Anthropic | 23.64% |
| 27 | [GPT-5.6 Terra](/models/gpt-5-6-terra) | Codex · max reasoning | OpenAI | 21.52% |
| 28 | [Grok 4.6](/models/grok-4-6) | Grok Build · high reasoning | xAI | 19.70% |
| 29 | [Gemini 3.8 Flash](/models/gemini-3-8-flash) | mini-SWE-agent · high reasoning | Google | 18.18% |
| 30 | [GPT-5.6 Luna](/models/gpt-5-6-luna) | Codex · max reasoning | OpenAI | 17.58% |
| 31 | [GPT-6 Luna](/models/gpt-6-luna) | Codex · max reasoning | OpenAI | 15.15% |
| 32 | [Muse Spark 1.3](/models/muse-spark-1-3) | Muse Code · xhigh reasoning | Meta | 14.85% |
| 33 | [Gemini 3.7 Flash](/models/gemini-3-7-flash) | mini-SWE-agent · high reasoning | Google | 12.42% |
| 34 | [Grok 4.5](/models/grok-4-5) | Grok Build · high reasoning | xAI | 12.42% |
| 35 | [Claude Sonnet 5](/models/claude-sonnet-5) | Claude Code · max reasoning | Anthropic | 12.12% |

## Terminal-Bench 4.0

Source snapshot: November 10, 2026 snapshot.

We mirror the official 4.0 table. The raw results compare complete model-and-agent systems. The latest normalized scores replace 3.0 within one external agentic benchmark family and cannot independently make a category ranking-eligible.

- [Terminal-Bench 4.0 leaderboard](https://hub.harborframework.com/datasets/terminal-bench/terminal-bench/latest?tab=leaderboard&leaderboard=4-0-0)
- [Launch and methodology](https://www.tbench.ai/news/terminal-bench-4-0)
- [Version 4.0.0 release](https://github.com/harbor-framework/terminal-bench/releases/tag/v4.0.0)
- [GitHub repository](https://github.com/harbor-framework/terminal-bench)

| Source model | Configuration | Tasks completed |
|---|---|---|
| [Claude Opus 5.5](https://www.anthropic.com/claude-opus-5-5) | max | 64.85% |
| [Claude Sonnet 5.5](https://www.anthropic.com/claude-sonnet-5-5) | max | 61.82% |
| [GPT-6 Astra](https://developers.openai.com/api/docs/models/gpt-6-astra) | max | 58.18% |
| [GPT-6.1 Sol](https://developers.openai.com/api/docs/models/gpt-6.1-sol) | max | 58.18% |
| [Claude Fable 5.1](https://docs.anthropic.com/en/docs/about-claude/models/all-models) | max | 57.88% |
| [GPT-6 Astra](https://developers.openai.com/api/docs/models/gpt-6-astra) | xhigh | 57.88% |
| [GPT-6 Astra](https://developers.openai.com/api/docs/models/gpt-6-astra) | high | 57.88% |
| [Claude Fable 5.1](https://docs.anthropic.com/en/docs/about-claude/models/all-models) | xhigh | 57.88% |
| [Claude Fable 5.1](https://docs.anthropic.com/en/docs/about-claude/models/all-models) | high | 54.55% |
| [GPT-6 Astra](https://developers.openai.com/api/docs/models/gpt-6-astra) | medium | 54.24% |
| [Claude Fable 5.1](https://docs.anthropic.com/en/docs/about-claude/models/all-models) | medium | 53.94% |
| [Claude Opus 5](https://www.anthropic.com/news/claude-opus-5) | xhigh | 53.94% |
| [Claude Opus 5](https://www.anthropic.com/news/claude-opus-5) | max | 51.82% |
| [GPT-6 Astra](https://developers.openai.com/api/docs/models/gpt-6-astra) | low | 50.61% |
| [Claude Opus 5](https://www.anthropic.com/news/claude-opus-5) | high | 50.30% |
| [GPT-6 Sol](https://developers.openai.com/api/docs/models/gpt-6-sol) | max | 49.39% |
| [Claude Opus 5](https://www.anthropic.com/news/claude-opus-5) | medium | 44.85% |
| [Claude Fable 5](https://www.anthropic.com/news/claude-fable-5-mythos-5) | max | 44.55% |
| [Claude Fable 5.1](https://docs.anthropic.com/en/docs/about-claude/models/all-models) | low | 43.33% |
| [GLM-5.3](https://docs.z.ai/guides/llm/glm-5.3) | max | 41.82% |
| [Grok 4.7](https://docs.x.ai/developers/models/grok-4.7) | xhigh | 37.58% |
| [GPT-5.6 Sol](https://developers.openai.com/api/docs/models/gpt-5.6-sol) | max | 37.27% |
| [GLM-5.3-Flash](https://z.ai/blog/glm-5.3-flash) | none | 35.76% |
| [Claude Opus 5](https://www.anthropic.com/news/claude-opus-5) | low | 34.85% |
| [Qwen3.8-Max-0902](https://www.alibabacloud.com/en/notice/model_studio_update_notice_for_qwen38max_models_863) | max | 26.97% |
| [Claude Opus 4.8](https://www.anthropic.com/news/claude-opus-4-8) | max | 23.64% |
| [GPT-5.6 Terra](https://developers.openai.com/api/docs/models/gpt-5.6-terra) | max | 21.52% |
| [Grok 4.6](https://docs.x.ai/developers/models/grok-4.6) | high | 20.30% |
| [Gemini 3.8 Flash](https://ai.google.dev/gemini-api/docs/models/gemini-3.8-flash) | high | 19.09% |
| [GPT-5.6 Luna](https://developers.openai.com/api/docs/models/gpt-5.6-luna) | max | 17.27% |
| [GPT-6 Luna](https://developers.openai.com/api/docs/models/gpt-6-luna) | max | 16.36% |
| [Muse Spark 1.3](https://ai.meta.com/tools/muse/) | xhigh | 14.55% |
| [Grok 4.5](https://docs.x.ai/developers/models/grok-4.5) | high | 12.42% |
| [Claude Sonnet 5](https://www.anthropic.com/news/claude-sonnet-5) | max | 12.42% |
| [Gemini 3.7 Flash](https://ai.google.dev/gemini-api/docs/models/gemini-3.7-flash) | high | 11.21% |

## Terminal-Bench 4.0 (Vals)

Evaluation results published by [Vals AI](https://www.vals.ai/benchmarks/terminal-bench-4). Published leaderboard aggregates and evaluation settings only. Dataset contents are not included.

Source snapshot: November 9, 2026.

We keep the Vals-hosted Terminal-Bench 4.0 table on a dedicated key so it never overwrites the official Terminal-Bench 4.0 leaderboard rows. Vals runs its own harness and publishes per-domain splits; this mirror is display only and does not enter weighted rankings.

- [Vals Terminal-Bench 4.0 mirror on Vals AI](https://www.vals.ai/benchmarks/terminal-bench-4)
- [Vals methodology](https://www.vals.ai/methodology)
- [Vals home](https://www.vals.ai/home)

| Source model | Configuration | Vals Terminal-Bench 4.0 mirror score |
|---|---|---|
| [Claude Fable 5.2](https://www.vals.ai/models/anthropic_claude-fable-5-2) | Not reported | 65.15% |
| [Claude Opus 5.5](https://www.vals.ai/models/anthropic_claude-opus-5-5) | Not reported | 65.15% |
| [Claude Sonnet 5.5](https://www.vals.ai/models/anthropic_claude-sonnet-5-5) | Not reported | 64.14% |
| [Kimi K3.1 Pro](https://www.vals.ai/models/kimi_kimi-k3.1-pro) | max | 63.64% |
| [Gemini 4 Carbon](https://www.vals.ai/models/google_gemini-4-carbon) | high | 63.13% |
| [GPT-6 Astra](https://www.vals.ai/models/openai_gpt-6-astra) | max | 59.60% |
| [Claude Fable 5.1](https://www.vals.ai/models/anthropic_claude-fable-5-1) | Not reported | 58.08% |
| [Muse Wonder Full](https://www.vals.ai/models/meta_muse-wonder-full) | max | 58.08% |
| [Gemini 4 Argon](https://www.vals.ai/models/google_gemini-4-argon) | high | 57.58% |
| [GPT-6.1 Sol](https://www.vals.ai/models/openai_gpt-6.1-sol) | max | 55.05% |
| [Claude Opus 5](https://www.vals.ai/models/anthropic_claude-opus-5) | Not reported | 53.53% |
| [GPT-6.1 Terra](https://www.vals.ai/models/openai_gpt-6.1-terra) | max | 49.49% |
| [Muse Wonder](https://www.vals.ai/models/meta_muse-wonder) | max | 49.49% |
| [GPT-6 Sol](https://www.vals.ai/models/openai_gpt-6-sol) | max | 44.44% |
| [Claude Fable 5](https://www.vals.ai/models/anthropic_claude-fable-5) | Not reported | 41.41% |
| [Muse Ember](https://www.vals.ai/models/meta_muse-ember) | max | 39.39% |
| [Hy4](https://www.vals.ai/models/tencent_hy4) | Not reported | 39.39% |
| [GLM 5.3](https://www.vals.ai/models/zai_glm-5.3) | max | 38.89% |
| [Qwen3.8 Max Internal](https://www.vals.ai/models/alibaba_qwen3.8-max-internal) | Not reported | 38.89% |
| [GPT-5.6 Sol](https://www.vals.ai/models/openai_gpt-5.6-sol) | max | 37.88% |
| [GLM 5.3 Flash](https://www.vals.ai/models/zai_glm-5.3-flash) | max | 37.37% |
| [Claude Haiku 5.5](https://www.vals.ai/models/anthropic_claude-haiku-5-5) | Not reported | 36.36% |
| [Qwen3.8 Max](https://www.vals.ai/models/alibaba_qwen3.8-max) | Not reported | 34.34% |
| [DeepSeek V4.1 Flash 1017](https://www.vals.ai/models/deepseek_deepseek-v4.1-flash-1017) | max | 31.82% |
| [Step 5 Preview](https://www.vals.ai/models/stepfun_step-5-preview) | Not reported | 31.82% |
| [Mimo V2.6 Pro](https://www.vals.ai/models/xiaomi_mimo-v2.6-pro) | Not reported | 31.31% |
| [Grok 4.7](https://www.vals.ai/models/grok_grok-4.7) | xhigh | 28.79% |
| [Mistral Large 4](https://www.vals.ai/models/mistralai_mistral-large-4) | high | 27.78% |
| [Mistral Large 4 Oss](https://www.vals.ai/models/mistralai_mistral-large-4-oss) | high | 26.26% |
| [MiniMax M3.1 Flash](https://www.vals.ai/models/minimax_MiniMax-M3.1-flash) | Not reported | 25.25% |
| [Muse Spark 1.3 Max](https://www.vals.ai/models/meta_muse_spark_1_3_max) | max | 24.75% |
| [Mimo V2.6 Flash](https://www.vals.ai/models/xiaomi_mimo-v2.6-flash) | Not reported | 24.24% |
| [Claude Opus 4.8](https://www.vals.ai/models/anthropic_claude-opus-4-8) | Not reported | 23.23% |
| [GPT-5.6 Terra](https://www.vals.ai/models/openai_gpt-5.6-terra) | max | 22.73% |
| [Ember 1](https://www.vals.ai/models/fireworks_ember-1) | Not reported | 20.71% |
| [DeepSeek V4.1 Flash](https://www.vals.ai/models/deepseek_deepseek-v4.1-flash) | high | 19.70% |
| [Gemini 3.8 Flash](https://www.vals.ai/models/google_gemini-3.8-flash) | high | 19.70% |
| [DeepSeek V4 Flash 0731](https://www.vals.ai/models/deepseek_deepseek-v4-flash-0731) | high | 18.69% |
| [Grok 4.6](https://www.vals.ai/models/grok_grok-4.6) | high | 17.17% |
| [Kimi K3](https://www.vals.ai/models/kimi_kimi-k3) | max | 17.17% |
| [DeepSeek V4 Pro 0813](https://www.vals.ai/models/deepseek_deepseek-v4-pro-0813) | max | 14.14% |
| [GPT-6 Luna](https://www.vals.ai/models/openai_gpt-6-luna) | max | 13.64% |
| [Inkling 2 Small](https://www.vals.ai/models/thinkingmachines_inkling-2-small) | 0.99 | 12.63% |
| [Gemini 3.7 Flash](https://www.vals.ai/models/google_gemini-3.7-flash) | high | 12.12% |
| [GPT-5.6 Luna](https://www.vals.ai/models/openai_gpt-5.6-luna) | max | 11.62% |
| [DeepSeek V4 Pro](https://www.vals.ai/models/deepseek_deepseek-v4-pro) | max | 11.11% |
| [Claude Sonnet 5](https://www.vals.ai/models/anthropic_claude-sonnet-5) | Not reported | 10.61% |
| [Muse Spark 1.3](https://www.vals.ai/models/meta_muse_spark_1_3) | xhigh | 10.61% |
| [Grok 4.5](https://www.vals.ai/models/grok_grok-4.5) | high | 8.59% |
| [Hy4 Preview](https://www.vals.ai/models/tencent_hy4-preview) | Not reported | 8.08% |
| [Muse Spark Open](https://www.vals.ai/models/meta_muse-spark-open) | xhigh | 6.57% |
| [Gemini 3.5 Flash](https://www.vals.ai/models/google_gemini-3.5-flash) | high | 6.06% |
| [Muse Spark 1.2](https://www.vals.ai/models/meta_muse_spark_1_2) | xhigh | 6.06% |
| [Beam 1.0](https://www.vals.ai/models/reflection_Beam-1.0) | Not reported | 4.04% |
| [Mercury 3](https://www.vals.ai/models/inception_mercury-3) | high | 3.03% |
| [Gemini 3.1 Pro Preview](https://www.vals.ai/models/google_gemini-3.1-pro-preview) | high | 2.52% |
| [GPT-5.4 Mini](https://www.vals.ai/models/openai_gpt-5.4-mini-2026-03-17) | xhigh | 2.52% |
| [MiniMax M3](https://www.vals.ai/models/minimax_MiniMax-M3) | Not reported | 1.01% |
| [Inkling](https://www.vals.ai/models/thinkingmachines_inkling) | 0.99 | 0.51% |
| [Mimo V2.5 Pro](https://www.vals.ai/models/xiaomi_mimo-v2.5-pro) | Not reported | 0.51% |
| [Inkling Small](https://www.vals.ai/models/thinkingmachines_inkling-small) | 0.99 | 0.00% |
| [Nemotron 3.5 Super VL 120b A12b](https://www.vals.ai/models/baseten_nemotron-3.5-super-vl-120b-a12b) | high | 0.00% |
| [Mercury 2.5](https://www.vals.ai/models/inception_mercury-2.5) | high | 0.00% |

## Terminal-Bench version history

Every Terminal-Bench release, newest first. Scores are not directly comparable across releases that change the task set or resources (2.x, 3.0 and 4.0). Version 4.1 is a regrade of the 4.0 trials, so its scores stay within each entry’s 4.0 confidence interval.

| Release | Released | Tasks | Leader | What changed |
|---|---|---|---|---|
| [Terminal-Bench 4.1](https://www.tbench.ai/news/terminal-bench-4-1) | 2026-11-09 | 66 tasks | Claude Opus 5.5 (65.8%) | Regrades the saved 4.0 trials with verifiers hardened against reward hacking; no agent-facing changes. |
| [Terminal-Bench 4.0](https://www.tbench.ai/news/terminal-bench-4-0) | 2026-08-28 | 66 tasks | Claude Opus 5.5 (64.8%) | Recalibrates task resources, fixes 19 tasks and removes 8; fresh trials, not comparable with 3.0. |
| [Terminal-Bench 3.0](https://www.tbench.ai/news/terminal-bench-3-0) | 2026-07-30 | 74 tasks, 7 domains | Claude Opus 5 (42.7%) | New professional computer-work task set, first published as FrontierBench. |
| [Terminal-Bench 2.1](https://www.tbench.ai/news/terminal-bench-2-1) | 2026-05-06 | 89 tasks | SWE-2 (92.8%) | Fixes 28 Terminal-Bench 2.0 tasks. No official leaderboard; results are provider reports. |
| [Terminal-Bench 2.0](https://www.tbench.ai/news/announcement-2-0) | 2025-11-07 | 89 tasks | GPT-5.5 (84.7%) | Harder terminal task set released with the Harbor evaluation framework. |

## Model rank by Terminal-Bench version

Rank among models on each official board (best configuration per model).

| Model | 4.1 rank | 4.0 rank | 3.0 rank | 2.0 rank |
|---|---|---|---|---|
| Claude Opus 5.5 | #1 of 23 (65.8%) | #1 of 23 (64.8%) | — | — |
| Claude Sonnet 5.5 | #2 of 23 (63.9%) | #2 of 23 (61.8%) | — | — |
| GPT-6.1 Sol | #3 of 23 (60.9%) | #3 of 23 (58.2%) | — | — |
| GPT-6 Astra | #4 of 23 (60.3%) | #3 of 23 (58.2%) | — | — |
| Claude Fable 5.1 | #5 of 23 (57.6%) | #5 of 23 (57.9%) | — | — |
| Claude Opus 5 | #6 of 23 (53.3%) | #6 of 23 (53.9%) | #1 of 12 (42.7%) | — |
| GPT-6 Sol | #7 of 23 (50.3%) | #7 of 23 (49.4%) | — | — |
| Claude Fable 5 | #8 of 23 (44.9%) | #8 of 23 (44.5%) | #3 of 12 (34.0%) | — |
| GLM-5.3 | #9 of 23 (41.2%) | #9 of 23 (41.8%) | #4 of 12 (32.4%) | — |
| Grok 4.7 | #10 of 23 (38.8%) | #10 of 23 (37.6%) | — | — |
| GPT-5.6 Sol | #11 of 23 (38.5%) | #11 of 23 (37.3%) | #2 of 12 (34.6%) | — |
| GLM-5.3-Flash | #12 of 23 (35.5%) | #12 of 23 (35.8%) | — | — |
| Qwen3.8-Max-0902 | #13 of 23 (27.0%) | #13 of 23 (27.0%) | — | — |
| Claude Opus 4.8 | #14 of 23 (23.6%) | #14 of 23 (23.6%) | #6 of 12 (21.1%) | — |
| GPT-5.6 Terra | #15 of 23 (21.5%) | #15 of 23 (21.5%) | #7 of 12 (20.8%) | — |
| Grok 4.6 | #16 of 23 (19.7%) | #16 of 23 (20.3%) | #5 of 12 (26.5%) | — |
| Gemini 3.8 Flash | #17 of 23 (18.2%) | #17 of 23 (19.1%) | — | — |
| GPT-5.6 Luna | #18 of 23 (17.6%) | #18 of 23 (17.3%) | #11 of 12 (14.3%) | — |
| GPT-6 Luna | #19 of 23 (15.2%) | #19 of 23 (16.4%) | — | — |
| Muse Spark 1.3 | #20 of 23 (14.8%) | #20 of 23 (14.6%) | — | — |
| Grok 4.5 | #21 of 23 (12.4%) | #21 of 23 (12.4%) | #9 of 12 (15.7%) | — |
| Gemini 3.7 Flash | #21 of 23 (12.4%) | #23 of 23 (11.2%) | — | — |
| Claude Sonnet 5 | #23 of 23 (12.1%) | #21 of 23 (12.4%) | #10 of 12 (14.6%) | — |
| SWE-1.7 Lightning | — | — | #8 of 12 (18.6%) | — |
| GLM-5.2 | — | — | #12 of 12 (4.6%) | — |

## Terminal-Bench 3.0

| Rank | Model | Score |
|---|---|---|
| 1 | [Claude Opus 5](/models/claude-opus-5) | 42.7% |
| 2 | [GPT-5.6 Sol](/models/gpt-5-6-sol) | 34.6% |
| 3 | [Claude Fable 5](/models/claude-fable) | 34.0% |
| 4 | [GLM-5.3](/models/glm-5-3) | 32.4% |
| 5 | [Grok 4.6](/models/grok-4-6) | 26.5% |
| 6 | [Claude Opus 4.8](/models/claude-opus-4-8) | 21.1% |
| 7 | [GPT-5.6 Terra](/models/gpt-5-6-terra) | 20.8% |
| 8 | SWE-1.7 Lightning | 18.6% |
| 9 | [Grok 4.5](/models/grok-4-5) | 15.7% |
| 10 | [Claude Sonnet 5](/models/claude-sonnet-5) | 14.6% |
| 11 | [GPT-5.6 Luna](/models/gpt-5-6-luna) | 14.3% |
| 12 | [GLM-5.2](/models/glm-5-2) | 4.6% |

## Terminal-Bench 2.1

| Rank | Model | Score |
|---|---|---|
| 1 | [SWE-2](/models/swe-2) | 92.8% |
| 2 | [GPT-5.6 Sol](/models/gpt-5-6-sol) | 91.9% |
| 3 | [DeepSeek V4.1 Flash 1017](/models/deepseek-v4-1-flash) | 90.6% |
| 4 | [MiMo-V2.6-Pro](/models/mimo-v2-6-pro) | 89.9% |
| 5 | [Gemini 3.8 Flash](/models/gemini-3-8-flash) | 89.4% |
| 6 | [SWE-2 Flash](/models/swe-2-flash) | 89.2% |
| 7 | [Muse Spark 1.3](/models/muse-spark-1-3) | 88.8% |
| 8 | [LongCat-2.5](/models/longcat-2-5) | 88.4% |
| 9 | [Kimi K3](/models/kimi-k3) | 88.3% |
| 10 | [GLM-5.3](/models/glm-5-3) | 88.2% |
| 11 | [Claude Mythos 5](/models/claude-mythos-5) | 88.0% |
| 12 | [DeepSeek V4 Pro 0813](/models/deepseek-v4-pro-0813) | 87.9% |
| 13 | [MiMo-V2.6-Flash](/models/mimo-v2-6-flash) | 87.6% |
| 14 | [GPT-5.6 Terra](/models/gpt-5-6-terra) | 87.4% |
| 15 | [Qwen3.8 Max](/models/qwen3-8-max) | 86.6% |
| 16 | [Ornith-1.5-397B](/models/ornith-1-5-397b) | 86.1% |
| 17 | [Gemini 3.7 Flash](/models/gemini-3-7-flash) | 85.8% |
| 18 | [Hy4 preview](/models/hy4-preview) | 85.4% |
| 19 | [Inkling 2 Small](/models/inkling-2-small) | 85.1% |
| 20 | [Step 5 Preview](/models/step-5-preview) | 85.0% |
| 21 | [GPT-5.6 Luna](/models/gpt-5-6-luna) | 84.7% |
| 22 | [Claude Fable 5](/models/claude-fable) | 84.3% |
| 23 | [GLM-5.3-Flash](/models/glm-5-3-flash) | 84.3% |
| 24 | [Grok 4.5](/models/grok-4-5) | 83.3% |
| 25 | [Ultramarine 1 Flash](/models/ultramarine-1-flash) | 83.2% |
| 26 | [Muse Spark 1.2](/models/muse-spark-1-2) | 82.9% |
| 27 | [DeepSeek V4 Flash 0731](/models/deepseek-v4-flash-0731) | 82.7% |
| 28 | [Sakana Fugu-Ultra](/models/sakana-fugu-ultra) | 82.1% |
| 29 | [Ember-1](/models/ember-1) | 82.0% |
| 30 | [SWE-1.7](/models/swe-1-7) | 81.5% |
| 31 | [Ling 3.1 Flash](/models/ling-3-1-flash) | 81.2% |
| 32 | [GLM-5.2](/models/glm-5-2) | 81.0% |
| 33 | [Ripple-65B Preview](/models/ripple-65b-preview) | 80.5% |
| 34 | [Claude Sonnet 5](/models/claude-sonnet-5) | 80.4% |
| 35 | [Sakana Fugu](/models/sakana-fugu) | 80.2% |
| 36 | [Beam 1.0 501B-A23B](/models/reflection-beam) | 80.1% |
| 37 | [Muse Spark 1.1](/models/muse-spark-1-1) | 80.0% |
| 38 | [Particle-1](/models/particle-1) | 79.9% |
| 39 | [Beam 1.0 501B-A23B FP8](/models/reflection-beam-fp8) | 78.9% |
| 40 | [Beam 1.0 501B-A23B NVFP4](/models/reflection-beam-nvfp4) | 78.4% |
| 41 | [Atria Dawn Preview](/models/atria-dawn-preview) | 78.3% |
| 42 | [Ornith-1.0-397B](/models/ornith-1-0-397b) | 77.5% |
| 43 | [Gemini 3.5 Flash](/models/gemini-3-5-flash) | 76.2% |
| 44 | [dots3-note Preview](/models/dots3-note-preview) | 75.1% |
| 45 | [Claude Opus 4.8](/models/claude-opus-4-8) | 74.6% |
| 46 | [Jais-3-375B-A23B](/models/jais-3-375b-a23b) | 73.8% |
| 47 | [Qwen3.8-27B](/models/qwen3-8-27b) | 73.0% |
| 48 | [OrcaCyber-Nano-1.0](/models/orcacyber-nano-1-0) | 72.5% |
| 49 | [Seed 2.1 Pro](/models/seed-2-1-pro) | 71.0% |
| 50 | [Apodex 1.1](/models/apodex-1-1) | 70.8% |
| 51 | [Laguna S 2.1](/models/laguna-s-2-1) | 70.2% |
| 52 | [Quasar 438B](/models/quasar-438b) | 69.3% |
| 53 | [Ornith-1.5-35B-A3B](/models/ornith-1-5-35b-a3b) | 67.8% |
| 54 | [Seed 2.1 Turbo](/models/seed-2-1-turbo) | 67.6% |
| 55 | [Jais-3-MoVA-36B-A4B](/models/jais-3-mova-36b-a4b) | 67.3% |
| 56 | [Nemotron 3.5 Super VL 120B A12B](/models/nemotron-3-5-super-vl-120b-a12b) | 66.7% |
| 57 | [MiniMax M3](/models/minimax-m3) | 66.0% |
| 58 | [Pokee-Isaac 28B](/models/pokee-isaac-28b) | 65.1% |
| 59 | [Inkling-Small](/models/inkling-small) | 64.7% |
| 60 | [Ornith-1.0-35B](/models/ornith-1-0-35b) | 64.2% |
| 61 | [Inkling](/models/inkling) | 63.8% |
| 62 | [Jais-3-32B](/models/jais-3-32b) | 63.7% |
| 63 | [MAI-Code-1.1-Flash](/models/mai-code-1-1-flash) | 62.9% |
| 64 | [Nemotron 3.5 Super VL 120B A12B NVFP4](/models/nemotron-3-5-super-vl-120b-a12b-nvfp4) | 60.8% |
| 65 | [Step 3.7 Flash](/models/step-3-7-flash) | 59.5% |
| 66 | [Ling 3.0 Flash](/models/ling-3-0-flash) | 57.0% |
| 67 | [Solar Pro 4](/models/solar-pro-4) | 57.0% |
| 68 | [Nemotron 3 Ultra](/models/nemotron-3-ultra) | 56.4% |
| 69 | [Flint](/models/flint-40b-a2-5b) | 54.9% |
| 70 | [Gemini 3.5 Flash-Lite](/models/gemini-3-5-flash-lite) | 54.0% |
| 71 | [Nemotron 3.5 Bolt 74B-A9B BF16](/models/nemotron-3-5-bolt-74b-a9b-bf16) | 53.9% |
| 72 | [Ternary Bonsai 2 27B](/models/ternary-bonsai-2-27b) | 52.8% |
| 73 | [Nemotron 3.5 Bolt 74B-A9B](/models/nemotron-3-5-bolt-74b-a9b) | 51.0% |
| 74 | [Jais-3-7B](/models/jais-3-7b) | 46.4% |
| 75 | [Ornith-1.5-9B](/models/ornith-1-5-9b) | 46.2% |
| 76 | [K-EXAONE 2.0](/models/k-exaone-2-0) | 43.8% |
| 77 | [Ornith-1.0-9B](/models/ornith-1-0-9b) | 43.1% |
| 78 | [Jais-3-3.7B](/models/jais-3-3-7b) | 41.2% |
| 79 | [A.X K2](/models/a-x-k2) | 36.0% |
| 80 | [Granite 4.2 30B](/models/granite-4-2-30b) | 29.2% |
| 81 | [Nemotron 3.5 Lightning 30B A3B NVFP4](/models/nemotron-3-5-lightning-30b-a3b-nvfp4) | 23.5% |
| 82 | [Granite 4.2 8B](/models/granite-4-2-8b) | 20.6% |
| 83 | [Mellum2.1-12B-A2.5B-Thinking](/models/mellum2-1-12b-a2-5b-thinking) | 17.4% |
| 84 | [LFM3-1.7B](/models/lfm3-1-7b) | 12.9% |
| 85 | [MiniCPM5-2B](/models/minicpm5-2b) | 8.6% |

### Terminal-Bench 2.1 (Vals)

| Rank | Model | Score |
|---|---|---|
| 1 | [Claude Opus 5.5](/models/claude-opus-5-5) | 87.6% |
| 2 | [GPT-6 Astra](/models/gpt-6-astra) | 87.3% |
| 3 | [GPT-5.6 Sol](/models/gpt-5-6-sol) | 85.8% |
| 4 | [Claude Fable 5.1](/models/claude-fable-5-1) | 85.0% |
| 5 | [Claude Opus 5](/models/claude-opus-5) | 84.6% |
| 6 | [GPT-6 Sol](/models/gpt-6-sol) | 83.1% |
| 7 | [Claude Sonnet 5.5](/models/claude-sonnet-5-5) | 83.1% |
| 8 | [Gemini 3.8 Flash](/models/gemini-3-8-flash) | 81.3% |
| 9 | [Kimi K3](/models/kimi-k3) | 80.9% |
| 10 | [Claude Fable 5](/models/claude-fable) | 80.5% |
| 11 | [GPT-5.6 Luna](/models/gpt-5-6-luna) | 79.0% |
| 12 | Muse Spark 1.3 Max | 79.0% |
| 13 | [Grok 4.6](/models/grok-4-6) | 78.3% |
| 14 | [Gemini 3.7 Flash](/models/gemini-3-7-flash) | 77.5% |
| 15 | [GPT-5.6 Terra](/models/gpt-5-6-terra) | 77.5% |
| 16 | [GPT-5.5](/models/gpt-5-5) | 76.4% |
| 17 | [MiMo-V2.6-Flash](/models/mimo-v2-6-flash) | 76.4% |
| 18 | [DeepSeek V4.1 Flash 1017](/models/deepseek-v4-1-flash) | 74.5% |
| 19 | [Claude Sonnet 5](/models/claude-sonnet-5) | 74.5% |
| 20 | [Gemini 3.5 Flash](/models/gemini-3-5-flash) | 74.2% |
| 21 | [Gemini 3.6 Flash](/models/gemini-3-6-flash) | 73.8% |
| 22 | [Grok 4.7](/models/grok-4-7) | 73.4% |
| 23 | [GPT-6 Luna](/models/gpt-6-luna) | 73.0% |
| 24 | [Muse Spark 1.3](/models/muse-spark-1-3) | 72.3% |
| 25 | [Claude Opus 4.8](/models/claude-opus-4-8) | 71.9% |
| 26 | [GLM-5.3](/models/glm-5-3) | 71.5% |
| 27 | Gemini 3.1 Pro Preview | 70.8% |
| 28 | Claude Opus 4.8 Claude Code | 69.7% |
| 29 | [Muse Spark 1.2](/models/muse-spark-1-2) | 69.7% |
| 30 | [Muse Spark 1.1](/models/muse-spark-1-1) | 69.3% |
| 31 | [Claude Opus 4.7](/models/claude-opus-4-7) | 68.5% |
| 32 | [Grok 4.5](/models/grok-4-5) | 67.8% |
| 33 | [GLM-5.2](/models/glm-5-2) | 67.8% |
| 34 | [MiMo-V2.6-Pro](/models/mimo-v2-6-pro) | 67.8% |
| 35 | [Qwen3.8 Max](/models/qwen3-8-max) | 67.4% |
| 36 | [DeepSeek V4 Flash 0731](/models/deepseek-v4-flash-0731) | 67.0% |
| 37 | [Kimi K2.7 Code](/models/kimi-k2-7-code) | 67.0% |
| 38 | [GLM-5.3-Flash](/models/glm-5-3-flash) | 62.9% |
| 39 | [Qwen3.7 Max](/models/qwen3-7-max) | 61.0% |
| 40 | [MiMo-V2.5](/models/mimo-v2-5) | 60.7% |
| 41 | [Composer 2.5](/models/composer-2-5) | 58.4% |
| 42 | [Qwen3.8-27B](/models/qwen3-8-27b) | 58.4% |
| 43 | [Gemini 4 Argon](/models/gemini-4-argon) | 57.7% |
| 44 | GPT-5.5 Codex | 57.3% |
| 45 | GPT-5.5 Factory | 57.3% |
| 46 | [Claude Sonnet 4.6](/models/claude-sonnet-4-6) | 57.3% |
| 47 | [MiMo-V2.5-Pro](/models/mimo-v2-5-pro) | 57.3% |
| 48 | [GLM-5.1](/models/glm-5-1) | 56.9% |
| 49 | [Inkling-Small](/models/inkling-small) | 55.1% |
| 50 | [Hy4 preview](/models/hy4-preview) | 55.1% |
| 51 | [GPT-5.4 mini](/models/gpt-5-4-mini) | 54.7% |
| 52 | [DeepSeek V4 Pro 0813](/models/deepseek-v4-pro-0813) | 54.7% |
| 53 | Gemini 3 Flash Preview | 53.9% |
| 54 | [Kimi K2.6](/models/kimi-2-6) | 53.6% |
| 55 | [MiniMax M3](/models/minimax-m3) | 53.6% |
| 56 | [Qwen3.6 Plus](/models/qwen3-6-plus) | 53.2% |
| 57 | [Qwen3.7 Plus](/models/qwen3-7-plus) | 52.8% |
| 58 | Nemotron 3 Ultra 550b A55b | 50.9% |
| 59 | [Gemini 3.5 Flash-Lite](/models/gemini-3-5-flash-lite) | 50.2% |
| 60 | Ling 3.0 Flash 2607 | 50.2% |
| 61 | Ling 3.0 Flash Af Rc3 | 50.2% |
| 62 | [DeepSeek V4 Pro 0813](/models/deepseek-v4-pro-0813) | 50.2% |
| 63 | [MiniMax M2.7](/models/minimax-m2-7) | 48.7% |
| 64 | [Inkling](/models/inkling) | 47.6% |
| 65 | Grok 4.20 0309 Reasoning | 44.2% |
| 66 | [Claude Haiku 4.5 Thinking](/models/claude-haiku-4-5-thinking) | 43.8% |
| 67 | [Grok 4.3](/models/grok-4-3) | 41.9% |
| 68 | Kimi K2.5 Thinking | 41.9% |
| 69 | [GPT-5.4 nano](/models/gpt-5-4-nano) | 41.6% |
| 70 | Mistral Medium 3.5 | 39.0% |
| 71 | [Mercury 2.5](/models/mercury-2-5) | 34.5% |
| 72 | Gemini 3.1 Flash Lite Preview | 34.1% |
| 73 | [Laguna M.1](/models/laguna-m-1) | 34.1% |
| 74 | [Laguna XS.2](/models/laguna-xs-2) | 25.8% |
| 75 | Command A Plus 05 2026 | 17.6% |
| 76 | Nemotron Lightning 3p5 30b A3b | 10.9% |

## Terminal-Bench 2.0

| Rank | Model | Score |
|---|---|---|
| 1 | [GPT-5.5](/models/gpt-5-5) | 82% |
| 2 | [GPT-5.3 Codex](/models/gpt-5-3-codex) | 77.3% |
| 3 | [GPT-5.4](/models/gpt-5-4) | 75.1% |
| 4 | [Qwen3.7 Plus](/models/qwen3-7-plus) | 70.3% |
| 5 | [Qwen3.7 Max](/models/qwen3-7-max) | 69.7% |
| 6 | [Claude Opus 4.7 (Adaptive)](/models/claude-opus-4-7-adaptive) | 69.4% |
| 7 | [Composer 2.5](/models/composer-2-5) | 69.3% |
| 8 | [MiMo-V2.5-Pro](/models/mimo-v2-5-pro) | 68.4% |
| 9 | [DeepSeek V4 Pro 0813](/models/deepseek-v4-pro-0813) | 67.9% |
| 10 | [Kimi K2.6](/models/kimi-2-6) | 66.7% |
| 11 | [MiMo-V2.5](/models/mimo-v2-5) | 65.8% |
| 12 | [Claude Opus 4.6](/models/claude-opus-4-6) | 65.4% |
| 13 | [Qwen 3.6 Max (preview)](/models/qwen3-6-max-preview) | 65.4% |
| 14 | [GLM-5.1](/models/glm-5-1) | 63.5% |
| 15 | [Composer 2](/models/composer-2) | 61.7% |
| 16 | [Qwen3.6 Plus](/models/qwen3-6-plus) | 61.6% |
| 17 | [GPT-5.4 mini](/models/gpt-5-4-mini) | 60% |
| 18 | [Qwen3.6-27B](/models/qwen3-6-27b) | 59.3% |
| 19 | [Claude Opus 4.5](/models/claude-opus-4-5) | 59.3% |
| 20 | [Claude Sonnet 4.6](/models/claude-sonnet-4-6) | 59.1% |
| 21 | [Muse Spark](/models/muse-spark) | 59% |
| 22 | [MiniMax M2.7](/models/minimax-m2-7) | 57% |
| 23 | [DeepSeek V4 Flash 0731](/models/deepseek-v4-flash-0731) | 56.9% |
| 24 | [GLM-5](/models/glm-5) | 56.2% |
| 25 | [Hy3 Preview](/models/hy3-preview) | 54.4% |
| 26 | [Qwen3.5 397B](/models/qwen3-5-397b) | 52.5% |
| 27 | [Qwen3.6-35B-A3B](/models/qwen3-6-35b-a3b) | 51.5% |
| 28 | [Kimi K2.5 (Reasoning)](/models/kimi-k2-5-reasoning) | 50.8% |
| 29 | [Kimi K2.5](/models/kimi-k2-5) | 50.8% |
| 30 | [Claude Sonnet 4.5](/models/claude-sonnet-4-5) | 50% |
| 31 | [Qwen3.5-122B-A10B](/models/qwen3-5-122b-a10b) | 49.4% |
| 32 | [Grok 4.20](/models/grok-4-20-beta) | 47.1% |
| 33 | [GPT-5.4 nano](/models/gpt-5-4-nano) | 46.3% |
| 34 | [MAI-Thinking-1](/models/mai-thinking-1) | 46% |
| 35 | [Laguna M.1](/models/laguna-m-1) | 45.8% |
| 36 | [Qwen3.5-27B](/models/qwen3-5-27b) | 41.6% |
| 37 | [GLM-4.7](/models/glm-4-7) | 41% |
| 38 | [Qwen3.5-35B-A3B](/models/qwen3-5-35b-a3b) | 40.5% |
| 39 | [Laguna XS 2.1](/models/laguna-xs-2-1) | 37.5% |
| 40 | [Laguna XS.2](/models/laguna-xs-2) | 35.7% |
| 41 | [LongCat-Flash-Lite-Sparse](/models/longcat-flash-lite-sparse) | 33.7% |

### Vals Terminal-Bench 2.0 mirror

| Rank | Model | Score |
|---|---|---|
| 1 | [GPT-5.5](/models/gpt-5-5) | 73.20% |
| 2 | [Claude Opus 4.8](/models/claude-opus-4-8) | 70.04% |
| 3 | [Claude Opus 4.7](/models/claude-opus-4-7) | 68.54% |
| 4 | [Gemini 3.5 Flash](/models/gemini-3-5-flash) | 67.42% |
| 5 | Gemini 3.1 Pro Preview | 67.42% |
| 6 | [GPT-5.3 Codex](/models/gpt-5-3-codex) | 64.05% |
| 7 | [Muse Spark](/models/muse-spark) | 59.55% |
| 8 | [Claude Sonnet 4.6](/models/claude-sonnet-4-6) | 59.55% |
| 9 | [Qwen3.7 Max](/models/qwen3-7-max) | 59.18% |
| 10 | [Claude Opus 4.5](/models/claude-opus-4-5) | 58.43% |
| 11 | [Claude Opus 4.6 (Adaptive)](/models/claude-opus-4-6-thinking) | 58.43% |
| 12 | [GPT-5.4](/models/gpt-5-4) | 58.43% |
| 13 | [Kimi K2.6](/models/kimi-2-6) | 57.30% |
| 14 | [DeepSeek V4 Pro 0813](/models/deepseek-v4-pro-0813) | 56.18% |
| 15 | Gemini 3 Pro Preview | 55.06% |
| 16 | [Claude Opus 4.5 Thinking](/models/claude-opus-4-5-thinking) | 53.93% |
| 17 | [GLM-5.1](/models/glm-5-1) | 53.93% |
| 18 | Gemini 3 Flash Preview | 51.69% |
| 19 | [GPT-5.2](/models/gpt-5-2) | 51.69% |
| 20 | [Qwen 3.6 Max (preview)](/models/qwen3-6-max-preview) | 51.69% |
| 21 | GLM 5 Thinking | 49.44% |
| 22 | [MiniMax M2.7](/models/minimax-m2-7) | 47.19% |
| 23 | [MiniMax M3](/models/minimax-m3) | 46.07% |
| 24 | [GPT-5.4 mini](/models/gpt-5-4-mini) | 44.94% |
| 25 | [Qwen3.6 Plus](/models/qwen3-6-plus) | 44.94% |
| 26 | [GPT-5.1](/models/gpt-5-1) | 44.94% |
| 27 | [Qwen3.6-27B](/models/qwen3-6-27b) | 44.94% |
| 28 | [Grok 4.3](/models/grok-4-3) | 43.45% |
| 29 | [Claude Sonnet 4.5 Thinking](/models/claude-sonnet-4-5-thinking) | 41.57% |
| 30 | [MiniMax M2.5](/models/minimax-m2-5) | 41.57% |
| 31 | Qwen3.5 Plus Thinking | 41.57% |
| 32 | Grok 4.20 0309 Reasoning | 40.45% |
| 33 | Kimi K2.5 Thinking | 40.45% |
| 34 | [GPT-5.4 nano](/models/gpt-5-4-nano) | 39.89% |
| 35 | Gemma 4 31b It | 39.33% |
| 36 | [Claude Haiku 4.5 Thinking](/models/claude-haiku-4-5-thinking) | 38.20% |
| 37 | [GLM-4.7](/models/glm-4-7) | 38.20% |
| 38 | MiniMax M2.1 | 37.08% |
| 39 | GPT-5 | 37.08% |
| 40 | Kimi K2 Thinking | 37.08% |
| 41 | DeepSeek V3p2 Thinking | 35.95% |
| 42 | DeepSeek V3p2 | 34.83% |
| 43 | [Laguna M.1](/models/laguna-m-1) | 31.46% |
| 44 | [Gemini 2.5 Pro](/models/gemini-2-5-pro) | 30.34% |
| 45 | Mistral Medium 3.5 | 30.34% |
| 46 | [Grok 4 Fast (Reasoning)](/models/grok-4-fast-reasoning) | 29.21% |
| 47 | [Laguna XS.2](/models/laguna-xs-2) | 28.09% |
| 48 | [GLM-4.6](/models/glm-4-6) | 28.09% |
| 49 | Grok 4 0709 | 28.09% |
| 50 | [GPT-5 mini](/models/gpt-5-mini) | 26.97% |
| 51 | Moonshotai Kimi K2 Instruct | 25.84% |
| 52 | [Grok 4.1 Fast (Reasoning)](/models/grok-4-1-fast-reasoning) | 24.72% |
| 53 | Gemini 3.1 Flash Lite Preview | 24.72% |
| 54 | [Qwen3.5 Flash](/models/qwen3-5-flash) | 24.72% |
| 55 | [Qwen3 Max](/models/qwen3-max) | 24.72% |
| 56 | DeepSeek V3p1 | 22.47% |
| 57 | Gemini 2.5 Flash Preview 09 2025 Thinking | 21.35% |
| 58 | [Qwen3 Max](/models/qwen3-max) | 20.23% |
| 59 | [GPT-OSS 120B](/models/gpt-oss-120b) | 19.10% |
| 60 | Grok 4.1 Fast Non Reasoning | 17.98% |
| 61 | [Trinity-Large-Thinking](/models/trinity-large-thinking) | 17.98% |
| 62 | Mistral Small 2603 | 16.85% |
| 63 | [GPT-4.1](/models/gpt-4-1) | 14.61% |
| 64 | Magistral Medium 2509 | 13.48% |
| 65 | Mistral Large 2512 | 8.99% |
| 66 | Command A 03 2025 | 2.25% |
| 67 | Llama4 Maverick Instruct Basic | 2.25% |

## FAQ

### What does Terminal-Bench 4.1 measure?

The current Terminal-Bench release keeps the 66 professional computer-work tasks of 4.0 and regrades every saved trial with verifiers hardened against reward hacking and overly strict checks.

### Which model leads the published Terminal-Bench 4.1 snapshot?

Claude Opus 5.5 currently leads the published Terminal-Bench 4.1 snapshot with a score of 65.76%.

### How many models are evaluated on Terminal-Bench 4.1?

The November 10, 2026 snapshot contains 35 AI models.

### Does Terminal-Bench 4.1 affect BenchLM's overall score?

Not directly. Terminal-Bench 4.1 is still displayed on BenchLM for reference, but it is excluded from the weighted scoring formula.