BlueBench
Agents solve enterprise tasks across knowledge, coding, security and business workflows. The benchmark aggregates capability scores from multiple public task suites.
IBM Research publishes an Apache 2.0 leaderboard space and describes BlueBench as an open benchmark developed with domain experts for enterprise use cases.
- Made by
- IBM Research
- Released
- September 2025
- Kind
- General assistants
- Level
- Expert
- Scored
- Checked automatically
- Answers
- Agent task
- Tasks
- All public
- Measures
- Capability
- Language
- English
- Interaction
- Agentic
- Environment
- APIs and tools
- Items from
- Written by experts
- Refresh
- One fixed set
- Run by
- Anyone
- Occupations
- Computer and mathematical
- In views
- Agents, Law and finance
- Strongest model tested
- o4-mini #95 on the capability index
- Last model added
- April 2025
- Scores refreshed
- November 21st, 2026
GPT-4.1 leads with 64.6 among models released by April 2025 GPT-4.1 leads with 64.6 among models released by April 2025
Best score per entry from the benchmark's own leaderboard, read on November 21st, 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 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
- 64.6 GPT-4.1
- 64.1 GPT-4o
- 63.0 GPT-4.1 mini
- 61.6 meta-llama/llama-3-405b-instruct
- 60.5 mistralai/mistral-medium-2505
- 59.0
Llama 3.3 70B Instruct
- 58.6 o1
- 57.0 GPT-4.1 nano
- 55.5
Mistral Large
- 54.1 mistralai/mistral-small-3-1-24b-instruct-2503
- 54.0 o4-mini
- 53.9 o3-mini
- 49.7 Granite-3.0-8B-Instruct
- 49.3 Granite 3.3 8B Instruct
- 45.4
Pixtral 12B
- 44.3 meta-llama/llama-3-2-3b-instruct
- 41.0 Granite-3.0-2B-Instruct
- 31.8 meta-llama/llama-3-2-1b-instruct
- 1 GPT-4.1 64.6
- 2 GPT-4o Mar 2025 64.1
- 3 GPT-4.1 mini 63.0
- 4 meta-llama/llama-3-405b-instruct 61.6
- 5 mistralai/mistral-medium-2505 60.5
- 6
Llama 3.3 70B Instruct 59.0
- 7 o1 58.6
- 8 GPT-4.1 nano 57.0
- 9
Mistral Large 55.5
- 10 mistralai/mistral-small-3-1-24b-instruct-2503 54.1
- 11 o4-mini 54.0
- 12 o3-mini 53.9
- 13 Granite-3.0-8B-Instruct 49.7
- 14 Granite 3.3 8B Instruct 49.3
- 15
Pixtral 12B 45.4
- 16 meta-llama/llama-3-2-3b-instruct 44.3
- 17 Granite-3.0-2B-Instruct 41.0
- 18 meta-llama/llama-3-2-1b-instruct 31.8
Source: Official BlueBench leaderboard
- 1 GPT-4.1 64.6
- 2 GPT-4o Mar 2025 64.1
- 3 GPT-4.1 mini 63.0
- 4
Llama 3.3 70B Instruct 59.0
- 5 o1 58.6
- 6 GPT-4.1 nano 57.0
- 7
Mistral Large 55.5
- 8 o4-mini 54.0
- 9 o3-mini 53.9
- 10 Granite-3.0-8B-Instruct 49.7
- 11 Granite 3.3 8B Instruct 49.3
- 12
Pixtral 12B 45.4
- 13 Granite-3.0-2B-Instruct 41.0
Source: Official BlueBench leaderboard
| # | Entry | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | GPT-4.1 | 64.6 | Apr 2025 | Official leaderboard | |
| 2 | GPT-4o Mar 2025 | 64.1 | Mar 2025 | Official leaderboard | |
| 3 | GPT-4.1 mini | 63.0 | Apr 2025 | Official leaderboard | |
| 4 | meta-llama/llama-3-405b-instruct | 61.6 | Official leaderboard | ||
| 5 | mistralai/mistral-medium-2505 | 60.5 | Official leaderboard | ||
| 6 | | 59.0 | Dec 2024 | Official leaderboard | |
| 7 | o1 | 58.6 | Dec 2024 | Official leaderboard | |
| 8 | GPT-4.1 nano | 57.0 | Apr 2025 | Official leaderboard | |
| 9 | | 55.5 | Feb 2024 | Official leaderboard | |
| 10 | mistralai/mistral-small-3-1-24b-instruct-2503 | 54.1 | Official leaderboard | ||
| 11 | o4-mini | 54.0 | Apr 2025 | Official leaderboard | |
| 12 | o3-mini | 53.9 | Jan 2025 | Official leaderboard | |
| 13 | Granite-3.0-8B-Instruct | 49.7 | Oct 2024 | Official leaderboard | |
| 14 | Granite 3.3 8B Instruct | 49.3 | Apr 2025 | Official leaderboard | |
| 15 | | 45.4 | Sep 2024 | Official leaderboard | |
| 16 | meta-llama/llama-3-2-3b-instruct | 44.3 | Official leaderboard | ||
| 17 | Granite-3.0-2B-Instruct | 41.0 | Oct 2024 | Official leaderboard | |
| 18 | meta-llama/llama-3-2-1b-instruct | 31.8 | Official leaderboard |
| # | Model | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | GPT-4.1 | 64.6 | Apr 2025 | Official leaderboard | |
| 2 | GPT-4o Mar 2025 | 64.1 | Mar 2025 | Official leaderboard | |
| 3 | GPT-4.1 mini | 63.0 | Apr 2025 | Official leaderboard | |
| 4 | | 59.0 | Dec 2024 | Official leaderboard | |
| 5 | o1 | 58.6 | Dec 2024 | Official leaderboard | |
| 6 | GPT-4.1 nano | 57.0 | Apr 2025 | Official leaderboard | |
| 7 | | 55.5 | Feb 2024 | Official leaderboard | |
| 8 | o4-mini | 54.0 | Apr 2025 | Official leaderboard | |
| 9 | o3-mini | 53.9 | Jan 2025 | Official leaderboard | |
| 10 | Granite-3.0-8B-Instruct | 49.7 | Oct 2024 | Official leaderboard | |
| 11 | Granite 3.3 8B Instruct | 49.3 | Apr 2025 | Official leaderboard | |
| 12 | | 45.4 | Sep 2024 | Official leaderboard | |
| 13 | Granite-3.0-2B-Instruct | 41.0 | Oct 2024 | Official leaderboard |
Rankings by part
We read 13 parts from the benchmark’s own leaderboard, each ranking models on one part of the test, such as Bias, Chatbot Abilities and Entity Extraction. o1 leads 3 of them, and 5 other models lead the rest. Arrows show how far a model moved from its overall rank among the models we track.
Bias
13 models we track have a score here.
- 1 o3-mini ↑8 98.0
- 2 GPT-4.1 ↓1 98.0
- 3
Mistral Large ↑4 97.0
- 4
Llama 3.3 70B Instruct 96.0
- 5 o1 94.9
- 6 GPT-4.1 mini ↓3 94.9
- 7 o4-mini ↑1 94.9
- 8 GPT-4o Mar 2025 ↓6 93.9
- 9 Granite-3.0-8B-Instruct ↑1 83.8
- 10 GPT-4.1 nano ↓4 81.8
Chatbot Abilities
13 models we track have a score here.
- 1 GPT-4.1 mini ↑2 97.5
- 2 GPT-4.1 ↓1 96.4
- 3 GPT-4.1 nano ↑3 95.6
- 4 GPT-4o Mar 2025 ↓2 94.1
- 5
Mistral Large ↑2 86.1
- 6
Pixtral 12B ↑6 76.5
- 7 Granite 3.3 8B Instruct ↑4 60.8
- 8 Granite-3.0-8B-Instruct ↑2 51.3
- 9
Llama 3.3 70B Instruct ↓5 49.7
- 10 o3-mini ↓1 37.1
Entity Extraction
13 models we track have a score here.
- 1 o1 ↑4 79.2
- 2 o4-mini ↑6 75.5
- 3 GPT-4o Mar 2025 ↓1 73.6
- 4 GPT-4.1 mini ↓1 73.2
- 5 o3-mini ↑4 72.9
- 6 GPT-4.1 ↓5 72.4
- 7
Llama 3.3 70B Instruct ↓3 70.2
- 8
Mistral Large ↓1 61.2
- 9 GPT-4.1 nano ↓3 61.1
- 10 Granite 3.3 8B Instruct ↑1 44.3
Knowledge
13 models we track have a score here.
- 1 o1 ↑4 71.4
- 2 GPT-4.1 ↓1 58.2
- 3
Mistral Large ↑4 55.1
- 4
Llama 3.3 70B Instruct 53.1
- 5 o3-mini ↑4 50.0
- 6 GPT-4o Mar 2025 ↓4 45.9
- 7 GPT-4.1 nano ↓1 44.9
- 8 Granite-3.0-8B-Instruct ↑2 43.9
- 9 GPT-4.1 mini ↓6 42.9
- 10 o4-mini ↓2 40.8
Legal
13 models we track have a score here.
- 1 GPT-4o Mar 2025 ↑1 60.9
- 2 GPT-4.1 ↓1 60.5
- 3
Llama 3.3 70B Instruct ↑1 59.5
- 4 GPT-4.1 mini ↓1 59.3
- 5 GPT-4.1 nano ↑1 56.9
- 6 o4-mini ↑2 56.4
- 7 Granite 3.3 8B Instruct ↑4 55.9
- 8 o1 ↓3 51.2
- 9 Granite-3.0-8B-Instruct ↑1 49.7
- 10 Granite-3.0-2B-Instruct ↑3 49.2
News Classification
13 models we track have a score here.
- 1 GPT-4o Mar 2025 ↑1 69.1
- 2 o1 ↑3 68.2
- 3 o4-mini ↑5 67.4
- 4 GPT-4.1 mini ↓1 66.7
- 5 GPT-4.1 ↓4 66.3
- 6
Llama 3.3 70B Instruct ↓2 59.9
- 7 Granite-3.0-8B-Instruct ↑3 50.0
- 8 Granite 3.3 8B Instruct ↑3 46.1
- 9 o3-mini 43.1
- 10 GPT-4.1 nano ↓4 42.2
Product Help
13 models we track have a score here.
- 1
Llama 3.3 70B Instruct ↑3 86.6
- 2 o3-mini ↑7 86.1
- 3 GPT-4o Mar 2025 ↓1 85.9
- 4 GPT-4.1 mini ↓1 84.9
- 5 o4-mini ↑3 84.3
- 6 o1 ↓1 83.0
- 7 GPT-4.1 ↓6 82.3
- 8 Granite-3.0-8B-Instruct ↑2 75.0
- 9 Granite 3.3 8B Instruct ↑2 73.9
- 10 Granite-3.0-2B-Instruct ↑3 70.7
QA Finance
13 models we track have a score here.
- 1 GPT-4o Mar 2025 ↑1 37.0
- 2 GPT-4.1 ↓1 33.0
- 3 GPT-4.1 mini 28.0
- 4 GPT-4.1 nano ↑2 26.0
- 5
Llama 3.3 70B Instruct ↓1 24.0
- 6 o3-mini ↑3 23.0
- 7 o1 ↓2 22.0
- 8 o4-mini 22.0
- 9
Mistral Large ↓2 19.0
- 10 Granite-3.0-8B-Instruct 16.0
Rag General
13 models we track have a score here.
- 1 GPT-4.1 mini ↑2 54.6
- 2 Granite-3.0-8B-Instruct ↑8 54.0
- 3 Granite 3.3 8B Instruct ↑8 51.9
- 4
Mistral Large ↑3 51.1
- 5 GPT-4.1 nano ↑1 50.9
- 6
Llama 3.3 70B Instruct ↓2 49.0
- 7 GPT-4o Mar 2025 ↓5 48.9
- 8 GPT-4.1 ↓7 48.5
- 9
Pixtral 12B ↑3 47.8
- 10 Granite-3.0-2B-Instruct ↑3 46.3
Reasoning
13 models we track have a score here.
- 1 o1 ↑4 79.0
- 2 GPT-4o Mar 2025 78.0
- 3 GPT-4.1 ↓2 77.5
- 4 o4-mini ↑4 75.5
- 5
Mistral Large ↑2 75.0
- 6
Pixtral 12B ↑6 73.5
- 7
Llama 3.3 70B Instruct ↓3 73.5
- 8 o3-mini ↑1 73.0
- 9 GPT-4.1 mini ↓6 73.0
- 10 GPT-4.1 nano ↓4 69.0
Safety
13 models we track have a score here.
- 1 Granite 3.3 8B Instruct ↑10 88.4
- 2
Mistral Large ↑5 88.0
- 3 Granite-3.0-8B-Instruct ↑7 88.0
- 4 GPT-4o Mar 2025 ↓2 87.8
- 5 GPT-4.1 ↓4 87.5
- 6 GPT-4.1 nano 86.7
- 7 GPT-4.1 mini ↓4 86.0
- 8 Granite-3.0-2B-Instruct ↑5 84.9
- 9
Llama 3.3 70B Instruct ↓5 83.8
- 10 o4-mini ↓2 83.5
Summarization
13 models we track have a score here.
- 1
Llama 3.3 70B Instruct ↑3 19.5
- 2
Mistral Large ↑5 18.2
- 3 GPT-4.1 ↓2 18.2
- 4 Granite-3.0-8B-Instruct ↑6 18.0
- 5 GPT-4.1 mini ↓2 17.9
- 6
Pixtral 12B ↑6 17.3
- 7 Granite-3.0-2B-Instruct ↑6 17.2
- 8 GPT-4o Mar 2025 ↓6 17.1
- 9 Granite 3.3 8B Instruct ↑2 17.1
- 10 GPT-4.1 nano ↓4 16.9
Translation
13 models we track have a score here.
- 1
Llama 3.3 70B Instruct ↑3 41.7
- 2 GPT-4o Mar 2025 41.4
- 3 GPT-4.1 ↓2 41.2
- 4 GPT-4.1 nano ↑2 40.3
- 5 GPT-4.1 mini ↓2 40.2
- 6
Mistral Large ↑1 39.8
- 7 o3-mini ↑2 37.8
- 8 o1 ↓3 37.4
- 9 o4-mini ↓1 33.4
- 10
Pixtral 12B ↑2 32.5
Number one over time
Lab claims count from the day the card came out, independent runs from the model’s release.
Source: Official BlueBench 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 BlueBench 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 Tool use
- AutomationBench In use Agent completes end-to-end business workflows inside a simulated company exposing dozens of REST APIs. Graded on resulting state.
Gemini 4 Neon 56.3%
- MCP Atlas In use Multi-step workflows over real MCP servers. Scores tool-use competency on realistic tasks.
Gemini 4 Argon 90.7%
- BFCL V4 In use Models call functions and tools, including multi-turn agentic tasks with web search and memory. Checked against expected calls and outcomes.
Qwen3.7-Max 75%
- τ³-bench Banking In use Agents act as banking customer-service reps over many turns with a simulated user, following policy. Pass^1 on the final database state.
Qwen 3.8 Max 55.1%
- Toolathlon-Verified In use 108 long-horizon tasks across 32 apps through about 600 MCP tools. Execution-based checkers inspect final environment state. GLM-5.3-Flash 78.4%
- DeepSearchQA In use Questions that need multi-step search and synthesis. Final answers graded against references. Kimi K3 95%
Data audit
8 of 9 checks pass, 1 warning. Look first at the Newest score check. The newest came 19 months ago, in April 2025, for Granite 3.3 8B Instruct. Every benchmark
Catalog record 2 pass
- Pass Task access All tasks are public.
- Pass Confidence Checked against a primary source.
Scores 1 warning 2 pass 2 notes
- Pass Independent runs Scores from the official leaderboard (18 entries).
- Pass Models scored 13 models we track, plus 5 other entries on its leaderboard.
- Warning Newest score The newest came 19 months ago, in April 2025, for Granite 3.3 8B Instruct.
- Note Frontier models None of the top 10 has a score. The strongest tested is o4-mini (#95).
- Note Headroom Scored as a plain number, with no fixed ceiling.
Leaderboard import 3 pass
- Pass Importer We read the board on huggingface.co: 13 rows for models we track and 5 other entries at the last read.
- Pass Last read November 21st, 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.