Log inOpen app
Mercor
318 posts
user avatar
Mercor
@mercor_ai
Organizing human intelligence to power the AI economy.
San Francisco
mercor.com/careers
Joined April 2021
27
Following
20.8K
Followers
user avatar
@mercor_ai

See Mercor’s full profile

Open X

New to X?

Sign up now to get your own personalized timeline!

Create account

By signing up, you agree to the Terms of Service and Privacy Policy, including Cookie Use.

Terms·Privacy·Cookies·Accessibility·Ads Info·© 2026 X Corp.
Don't miss what's happening
People on X are the first to know.
Log inSign up
  • Mercor reposted
    user avatar
    Brendan (can/do)
    Mercor
    @BrendanFoody
    10h
    Mercor grew from $1M to $2B in ARR in 24 months. This is the fastest growth trajectory in history. While everyone knows about AI lab investments in the Task Economy, our fastest-growing segment is Enterprise.
    user avatar
    Everett Randle
    Benchmark
    @EverettRandle
    10h
    Article cover image
    Article
    The Task Economy - Data will be the next $1 Trillion Category
    The Token Economy When we talk about AI today, tokens are king. Specifically, inference tokens have emerged as the primary proxy for tracking the growth of the AI ecosystem. Public companies report...
    20
    22
    291
    252K
  • Mercor reposted
    user avatar
    Everett Randle
    Benchmark
    @EverettRandle
    10h
    Article cover image
    Article
    The Task Economy - Data will be the next $1 Trillion Category
    The Token Economy When we talk about AI today, tokens are king. Specifically, inference tokens have emerged as the primary proxy for tracking the growth of the AI ecosystem. Public companies report...
    34
    41
    425
    351K
  • user avatar
    Mercor
    @mercor_ai
    10h
    Replying to @mercor_ai
    One caveat. Anthropic notes the new classifier flags more benign coding requests, and we see that in the shifted tool mix. Small output-type samples are noisy too (slides n=6). But on the metric that matters, whether or not AI models can do real professional work, the new
    1
    352
    user avatar
    Mercor
    @mercor_ai
    10h
    Download the APEX-Agents dataset on @huggingface: huggingface.co/datasets/merco… Open-source infra + eval service (Archipelago) on @github: github.com/Mercor-Intelli…
    mercor/apex-agents · Datasets at Hugging Face
    From huggingface.co
    1
    166
  • user avatar
    Mercor
    @mercor_ai
    10h
    Anthropic redeployed @claudeai Fable 5 with new cybersecurity safeguards, including a classifier that reroutes blocked requests to Opus 4.8. We re-ran the Fable 5 on APEX-Agents to see if the guardrails cost any capability. They didn't. Pass@1 43.3%, mean score 59.2%. Across
    APEX-Agents | Fable 5 re-release
    3
    1
    18
    24K
    user avatar
    Mercor
    @mercor_ai
    10h
    The re-release is also slightly leaner. It used fewer tokens on average (1.29M vs 1.37M) at nearly identical step counts (14.1 vs 14.3). Its tool mix shifted too: less code execution (22.5% vs 27.0%), more reading and inspection. It solved 237 tasks to the original's 234,
    APEX-Agents | Fable 5 re-release
    1
    397
  • Mercor reposted
    user avatar
    Brendan (can/do)
    Mercor
    @BrendanFoody
    Jul 6
    Mercor crossed $2B in ARR in June, just 4 months after hitting $1B in ARR. The civilization-scale effort to collect data is underway.
    user avatar
    will depue
    @willdepue
    Jul 6
    A Stargate for Data Labs are on a trajectory towards >$100B/year of data spend by 2030. As we begin the trillion-dollar compute project, we need to think about the equivalent civilizational-scale effort for the other core ingredient: data. At the foundation of the scaling
    57
    80
    987
    667K
  • Mercor reposted
    user avatar
    will depue
    @willdepue
    Jul 6
    A Stargate for Data Labs are on a trajectory towards >$100B/year of data spend by 2030. As we begin the trillion-dollar compute project, we need to think about the equivalent civilizational-scale effort for the other core ingredient: data. At the foundation of the scaling
    119
    143
    1.6K
    1.1M
  • user avatar
    Mercor
    @mercor_ai
    Jul 3
    Fable 5 is back and we’ve got results for the re-released version on APEX-SWE. While it did not perform as well as its earlier version from June, the model still significantly outperforms Opus 4.8. Fable 5 (June): 65.5% Pass@1 Fable 5 (July): 54.8% Pass@1 Opus 4.8: 45.3% Pass@1
    APEX-SWE | Fable 5 re-release
    73
    71
    1K
    245K
    user avatar
    Mercor
    @mercor_ai
    Jul 3
    APEX-SWE evaluates AI models across two different areas of software engineering work, Integration and Observability. Here is how the Fable 5 re-release performed in both areas compared to the June release. Integration Fable 5 (June): 61.33% Fable 5 (July): 59.33% Observability
    APEX-SWE | Fable 5 re-release
    4
    30
    9.2K
  • Mercor reposted
    user avatar
    Brendan (can/do)
    Mercor
    @BrendanFoody
    Jul 2
    Mercor now pays out over $4M / day to experts on our platform, with an average pay rate of over $100 / hour. Despite today's jobs report missing estimates, training agents is becoming one of the fastest-growing job categories in the world. Organizing human intelligence is the
    15
    24
    380
    55K
  • user avatar
    Mercor
    @mercor_ai
    Jul 1
    Replying to @mercor_ai
    Download the APEX-Agents dataset: huggingface.co/datasets/merco… Open-source infra + harness: github.com/Mercor-Intelli… Technical report: arxiv.org/pdf/2601.08806 APEX leaderboards: mercor.com/apex/
    mercor/APEX-SWE · Datasets at Hugging Face
    From huggingface.co
    1
    3
    1.2K
  • user avatar
    Mercor
    @mercor_ai
    Jul 1
    Replying to @mercor_ai
    Kimi K2.7 is now the second-best open-source model on APEX-SWE. Leaderboard Pass@1: 32.48%, 11th overall Integration Pass@1: 47.30% Observability Pass@1: 17.67% That's a strong improvement from Kimi K2.5, which scored 30.3% on the leaderboard and 41% on Integration. K2.7
    APEX-SWE | Kimi K2.7 Code
    1
    7
    1.5K
  • user avatar
    Mercor
    @mercor_ai
    Jul 1
    Replying to @mercor_ai
    GLM 5.2 is also the top open-source model on APEX-Agents. Leaderboard Pass@1: 35.6%, 5th overall Corporate Law: 29.9%, 5th Investment Banking: 38.8%, 5th Management Consulting: 38.1%, 13th Landing two domains in the top 5 on APEX-Agents is impressive for any model, open
    APEX-Agents | GLM-5.2
    1
    7
    1.3K
  • user avatar
    Mercor
    @mercor_ai
    Jul 1
    GLM 5.2 just became the first open-source model to lead a category on APEX-SWE. It scored a 55.3% Pass@1 on Integration, the top score we've recorded for any model, open or closed source. On the overall leaderboard, GLM 5.2 scored 37.3% Pass@1, ranking 6th place. That makes it
    APEX-SWE | GLM-5.2
    10
    35
    353
    23K
  • user avatar
    Mercor
    @mercor_ai
    Jul 1
    Claude Sonnet 5 is #10 on APEX-Agents, and it’s a clear generational jump. Compared to Sonnet 4.6: Pass@1: 23.7% → 32.5% (+8.8 pts) Mean score: 40.7% → 48.5% (+7.8 pts) @claudeai Sonnet 5 improved on the overall APEX-Agents leaderboard from being ranked #15 to #10 for Pass@1,
    APEX-Agents | Claude Sonnet 5
    12
    3
    43
    3.9K
    user avatar
    Mercor
    @mercor_ai
    Jul 1
    Two APEX-Agents domains demonstrated big improvements from Sonnet 4.6. For Investment Banking Analyst tasks, Sonnet 5 moved up from #18 to #8 (+11.2pts) mean score. For Management Consultant tasks, it improved from #19 to #13 (+12.3 pts) mean score. See the full APEX-Agents
    Mercor Logo
    APEX-Agents: AI Rankings for Agentic Tasks | Mercor
    From mercor.com
    2
    5
    717
    user avatar
    Mercor
    @mercor_ai
    Jul 1
    Subscribe to the APEX newsletter: mercor.com/apex/#newslett… Download the APEX-Agents dataset: huggingface.co/datasets/merco… Open-source infra + eval service (Archipelago): github.com/Mercor-Intelli… Technical report: arxiv.org/abs/2601.14242
    APEX - Mercor
    APEX Benchmarks: The AI Productivity Index | Mercor
    From mercor.com
    2
    4
    632
  • user avatar
    Mercor
    @mercor_ai
    Jun 30
    Our analysis on the performance of Sonnet 5 on APEX-SWE: @claudeai Sonnet 5 scored 43.7% Pass@1, behind Fable 5 (65.5%) and Opus 4.8 (45.3%). Our updated APEX-SWE leaderboard ranks Sonnet 5 in 3rd place. The model demonstrated major improvements in Integration and
    APEX-SWE | Claude Sonnet 5
    4
    5
    22
    2.5K
    user avatar
    Mercor
    @mercor_ai
    Jun 30
    Want to learn more about how Sonnet 5 and other models did on APEX-SWE? Sign up for the APEX newsletter: mercor.com/apex/#newslett… Download the APEX-SWE dataset: huggingface.co/datasets/merco… Eval harness: github.com/Mercor-Intelli… Technical report: arxiv.org/pdf/2601.08806
    756