Skip to main content

Shipped catalogue: 17 worker profiles across 5 execution harnesses.

Worker catalogue

Meet the workers behind every route.

A worker is the routable unit: usually a harness paired with a model, with provider-hosted exceptions. Each carries its own price and evidence, so the same CLI on another model is measured separately.

workers routable
99
harnesses · Claude Code, Codex, Cursor, Devin, OpenCode
5
models driven
84
attempts before a rate
20
workers ranked
0
nothing attributable yet
84

Browse workers

Compare workers and their evidence.

Search by worker, harness, or model. Rates stay withheld below 20 attributable attempts; no worker here has crossed that threshold yet.

No filters active

99 of 99 workers

Sorted by Harness, then model · public.worker_scoreboard · all workspaces · cached 5 minutes

  • Claude Code + Claude Fable 5Claude Code · anthropic/claude-fable-5
    Input / 1M
    $10.00
    Output / 1M
    $50.00

    Evidence

    No attributable attempts — nothing measured

  • Claude Code + Claude Haiku 4.5Claude Code · anthropic/claude-haiku-4.5
    Input / 1M
    $1.00
    Output / 1M
    $5.00

    Evidence

    No attributable attempts — nothing measured

  • Claude Code + Claude Opus 4.5Claude Code · anthropic/claude-opus-4.5
    Input / 1M
    $5.00
    Output / 1M
    $25.00

    Evidence

    No attributable attempts — nothing measured

  • Claude Code + Claude Opus 4.6Claude Code · anthropic/claude-opus-4.6
    Input / 1M
    $5.00
    Output / 1M
    $25.00

    Evidence

    No attributable attempts — nothing measured

  • Claude Code + Claude Opus 4.7Claude Code · anthropic/claude-opus-4.7
    Input / 1M
    $5.00
    Output / 1M
    $25.00

    Evidence

    No attributable attempts — nothing measured

  • Claude Code + Claude Opus 4.8Claude Code · anthropic/claude-opus-4.8
    Input / 1M
    $5.00
    Output / 1M
    $25.00

    Evidence

    No attributable attempts — nothing measured

  • Claude Code + OpusClaude Code · anthropic/claude-opus-5
    Input / 1M
    $5.00
    Output / 1M
    $25.00

    Evidence

    6 attempts — not yet enough to rank

  • Claude Code + Claude Opus 5 (Fast)Claude Code · anthropic/claude-opus-5-fast
    Input / 1M
    $10.00
    Output / 1M
    $50.00

    Evidence

    No attributable attempts — nothing measured

  • Claude Code + SonnetClaude Code · anthropic/claude-sonnet-5
    Input / 1M
    $2.00
    Output / 1M
    $10.00

    Evidence

    11 attempts — not yet enough to rank

  • Codex + GPT 5.6 LunaCodex · openai/gpt-5.6-luna
    Input / 1M
    $0.20
    Output / 1M
    $1.20

    Evidence

    No attributable attempts — nothing measured

  • Codex + GPT 5.6 Luna (Fast)Codex · openai/gpt-5.6-luna-fast
    Input / 1M
    $0.40
    Output / 1M
    $2.40

    Evidence

    No attributable attempts — nothing measured

  • Codex + GPTCodex · openai/gpt-5.6-sol
    Input / 1M
    $1.75
    Output / 1M
    $14.00

    Evidence

    11 attempts — not yet enough to rank

  • Codex + GPT 5.6 Sol (Fast)Codex · openai/gpt-5.6-sol-fast
    Input / 1M
    $4.00
    Output / 1M
    $20.00

    Evidence

    No attributable attempts — nothing measured

  • Codex + GPT 5.6 TerraCodex · openai/gpt-5.6-terra
    Input / 1M
    $2.00
    Output / 1M
    $12.00

    Evidence

    No attributable attempts — nothing measured

  • Codex + GPT 5.6 Terra (Fast)Codex · openai/gpt-5.6-terra-fast
    Input / 1M
    $4.00
    Output / 1M
    $24.00

    Evidence

    No attributable attempts — nothing measured

  • Cursor Cloud AgentCursor

    Model chosen by the provider

    Input / 1M
    Output / 1M

    Evidence

    No attributable attempts — nothing measured

  • DevinDevin

    Model chosen by the provider

    Input / 1M
    Output / 1M

    Evidence

    No attributable attempts — nothing measured

  • Qwen3.7 Flashalibaba/qwen3.7-flash
    Input / 1M
    $0.03
    Output / 1M
    $0.13

    Evidence

    No attributable attempts — nothing measured

  • Amazon Nova Liteamazon/nova-lite
    Input / 1M
    $0.06
    Output / 1M
    $0.24

    Evidence

    1 attempt — not yet enough to rank

  • Gemini 2.5 Flash Litegoogle/gemini-2.5-flash-lite
    Input / 1M
    $0.10
    Output / 1M
    $0.40

    Evidence

    No attributable attempts — nothing measured

  • Nemotron 3.5 Lightningnvidia/nemotron-3.5-lightning
    Input / 1M
    $0.05
    Output / 1M
    $0.20

    Evidence

    No attributable attempts — nothing measured

  • GPT-5 Nanoopenai/gpt-5-nano
    Input / 1M
    $0.05
    Output / 1M
    $0.40

    Evidence

    No attributable attempts — nothing measured

  • OpenCode + Qwen3-14BOpenCode · alibaba/qwen-3-14b
    Input / 1M
    $0.12
    Output / 1M
    $0.24

    Evidence

    No attributable attempts — nothing measured

  • OpenCode + Qwen3 235B A22BOpenCode · alibaba/qwen-3-235b
    Input / 1M
    $0.22
    Output / 1M
    $0.88

    Evidence

    No attributable attempts — nothing measured

  • OpenCode + Qwen3-30B-A3BOpenCode · alibaba/qwen-3-30b
    Input / 1M
    $0.12
    Output / 1M
    $0.50

    Evidence

    No attributable attempts — nothing measured

  • OpenCode + Qwen 3 32BOpenCode · alibaba/qwen-3-32b
    Input / 1M
    $0.16
    Output / 1M
    $0.64

    Evidence

    No attributable attempts — nothing measured

  • OpenCode + Qwen 3.6 Max PreviewOpenCode · alibaba/qwen-3.6-max-preview
    Input / 1M
    $1.30
    Output / 1M
    $7.80

    Evidence

    No attributable attempts — nothing measured

  • OpenCode + Qwen3 VL 235B A22B ThinkingOpenCode · alibaba/qwen3-235b-a22b-thinking
    Input / 1M
    $0.40
    Output / 1M
    $4.00

    Evidence

    No attributable attempts — nothing measured

  • OpenCode + Qwen3 Coder 480B A35B InstructOpenCode · alibaba/qwen3-coder
    Input / 1M
    $1.50
    Output / 1M
    $7.50

    Evidence

    No attributable attempts — nothing measured

  • OpenCode + Qwen 3 Coder 30B A3B InstructOpenCode · alibaba/qwen3-coder-30b-a3b
    Input / 1M
    $0.15
    Output / 1M
    $0.60

    Evidence

    No attributable attempts — nothing measured

  • OpenCode + Qwen3 Coder NextOpenCode · alibaba/qwen3-coder-next
    Input / 1M
    $0.50
    Output / 1M
    $1.20

    Evidence

    No attributable attempts — nothing measured

  • OpenCode + Qwen3 CoderOpenCode · alibaba/qwen3-coder-plus
    Input / 1M
    $1.00
    Output / 1M
    $5.00

    Evidence

    5 attempts — not yet enough to rank

  • OpenCode + Qwen3 MaxOpenCode · alibaba/qwen3-max
    Input / 1M
    $1.20
    Output / 1M
    $6.00

    Evidence

    No attributable attempts — nothing measured

  • OpenCode + Qwen3 Max PreviewOpenCode · alibaba/qwen3-max-preview
    Input / 1M
    $1.20
    Output / 1M
    $6.00

    Evidence

    No attributable attempts — nothing measured

  • OpenCode + Qwen 3 Max ThinkingOpenCode · alibaba/qwen3-max-thinking
    Input / 1M
    $1.20
    Output / 1M
    $6.00

    Evidence

    No attributable attempts — nothing measured

  • OpenCode + Qwen3 Next 80B A3B InstructOpenCode · alibaba/qwen3-next-80b-a3b-instruct
    Input / 1M
    $0.15
    Output / 1M
    $1.20

    Evidence

    No attributable attempts — nothing measured

  • OpenCode + Qwen3 Next 80B A3B ThinkingOpenCode · alibaba/qwen3-next-80b-a3b-thinking
    Input / 1M
    $0.15
    Output / 1M
    $1.20

    Evidence

    No attributable attempts — nothing measured

  • OpenCode + Qwen3 VL 235B A22B InstructOpenCode · alibaba/qwen3-vl-235b-a22b-instruct
    Input / 1M
    $0.40
    Output / 1M
    $1.60

    Evidence

    No attributable attempts — nothing measured

  • OpenCode + Qwen3 VL 235B A22B InstructOpenCode · alibaba/qwen3-vl-instruct
    Input / 1M
    $0.40
    Output / 1M
    $1.60

    Evidence

    No attributable attempts — nothing measured

  • OpenCode + Qwen3 VL 235B A22B ThinkingOpenCode · alibaba/qwen3-vl-thinking
    Input / 1M
    $0.40
    Output / 1M
    $4.00

    Evidence

    No attributable attempts — nothing measured

  • OpenCode + Qwen 3.5 FlashOpenCode · alibaba/qwen3.5-flash
    Input / 1M
    $0.10
    Output / 1M
    $0.40

    Evidence

    No attributable attempts — nothing measured

  • OpenCode + Qwen 3.5 PlusOpenCode · alibaba/qwen3.5-plus
    Input / 1M
    $0.40
    Output / 1M
    $2.40

    Evidence

    No attributable attempts — nothing measured

  • OpenCode + Qwen 3.6 27BOpenCode · alibaba/qwen3.6-27b
    Input / 1M
    $0.60
    Output / 1M
    $3.60

    Evidence

    No attributable attempts — nothing measured

  • OpenCode + Qwen 3.6 PlusOpenCode · alibaba/qwen3.6-plus
    Input / 1M
    $0.50
    Output / 1M
    $3.00

    Evidence

    No attributable attempts — nothing measured

  • OpenCode + Qwen 3.7 FlashOpenCode · alibaba/qwen3.7-flash
    Input / 1M
    $0.03
    Output / 1M
    $0.13

    Evidence

    No attributable attempts — nothing measured

  • OpenCode + Qwen 3.7 MaxOpenCode · alibaba/qwen3.7-max
    Input / 1M
    $2.50
    Output / 1M
    $7.50

    Evidence

    No attributable attempts — nothing measured

  • OpenCode + Qwen 3.7 PlusOpenCode · alibaba/qwen3.7-plus
    Input / 1M
    $0.40
    Output / 1M
    $1.60

    Evidence

    No attributable attempts — nothing measured

  • OpenCode + Qwen 3.8 MaxOpenCode · alibaba/qwen3.8-max
    Input / 1M
    $2.00
    Output / 1M
    $6.00

    Evidence

    No attributable attempts — nothing measured

  • OpenCode + Nova 2 LiteOpenCode · amazon/nova-2-lite
    Input / 1M
    $0.30
    Output / 1M
    $2.50

    Evidence

    No attributable attempts — nothing measured

  • OpenCode + Nova LiteOpenCode · amazon/nova-lite
    Input / 1M
    $0.06
    Output / 1M
    $0.24

    Evidence

    No attributable attempts — nothing measured

  • OpenCode + Nova MicroOpenCode · amazon/nova-micro
    Input / 1M
    $0.03
    Output / 1M
    $0.14

    Evidence

    No attributable attempts — nothing measured

  • OpenCode + Nova ProOpenCode · amazon/nova-pro
    Input / 1M
    $0.80
    Output / 1M
    $3.20

    Evidence

    No attributable attempts — nothing measured

  • OpenCode + Claude Fable 5OpenCode · anthropic/claude-fable-5
    Input / 1M
    $10.00
    Output / 1M
    $50.00

    Evidence

    No attributable attempts — nothing measured

  • OpenCode + Claude Haiku 4.5OpenCode · anthropic/claude-haiku-4.5
    Input / 1M
    $1.00
    Output / 1M
    $5.00

    Evidence

    No attributable attempts — nothing measured

  • OpenCode + Claude Opus 4OpenCode · anthropic/claude-opus-4
    Input / 1M
    $15.00
    Output / 1M
    $75.00

    Evidence

    No attributable attempts — nothing measured

  • OpenCode + Claude Opus 4.5OpenCode · anthropic/claude-opus-4.5
    Input / 1M
    $5.00
    Output / 1M
    $25.00

    Evidence

    No attributable attempts — nothing measured

  • OpenCode + Claude Opus 4.6OpenCode · anthropic/claude-opus-4.6
    Input / 1M
    $5.00
    Output / 1M
    $25.00

    Evidence

    No attributable attempts — nothing measured

  • OpenCode + Claude Opus 4.7OpenCode · anthropic/claude-opus-4.7
    Input / 1M
    $5.00
    Output / 1M
    $25.00

    Evidence

    No attributable attempts — nothing measured

  • OpenCode + Claude Opus 4.8OpenCode · anthropic/claude-opus-4.8
    Input / 1M
    $5.00
    Output / 1M
    $25.00

    Evidence

    No attributable attempts — nothing measured

  • OpenCode + Claude Opus 5OpenCode · anthropic/claude-opus-5
    Input / 1M
    $5.00
    Output / 1M
    $25.00

    Evidence

    No attributable attempts — nothing measured

  • OpenCode + ClaudeOpenCode · anthropic/claude-sonnet-5
    Input / 1M
    $2.00
    Output / 1M
    $10.00

    Evidence

    3 attempts — not yet enough to rank

  • OpenCode + Trinity Large ThinkingOpenCode · arcee-ai/trinity-large-thinking
    Input / 1M
    $0.25
    Output / 1M
    $0.90

    Evidence

    No attributable attempts — nothing measured

  • OpenCode + Seed 1.6OpenCode · bytedance/seed-1.6
    Input / 1M
    $0.25
    Output / 1M
    $2.00

    Evidence

    No attributable attempts — nothing measured

  • OpenCode + Bytedance Seed 1.8OpenCode · bytedance/seed-1.8
    Input / 1M
    $0.25
    Output / 1M
    $2.00

    Evidence

    No attributable attempts — nothing measured

  • OpenCode + Command AOpenCode · cohere/command-a
    Input / 1M
    $2.50
    Output / 1M
    $10.00

    Evidence

    No attributable attempts — nothing measured

  • OpenCode + DeepSeek R1OpenCode · deepseek/deepseek-r1
    Input / 1M
    $1.35
    Output / 1M
    $5.40

    Evidence

    3 attempts — not yet enough to rank

  • OpenCode + Gemini 2.5 ProOpenCode · google/gemini-2.5-pro
    Input / 1M
    $1.25
    Output / 1M
    $10.00

    Evidence

    1 attempt — not yet enough to rank

  • OpenCode + Gemini 3.5 Flash LiteOpenCode · google/gemini-3.5-flash-lite
    Input / 1M
    $0.30
    Output / 1M
    $2.50

    Evidence

    No attributable attempts — nothing measured

  • OpenCode + Gemini 3.6 FlashOpenCode · google/gemini-3.6-flash
    Input / 1M
    $0.75
    Output / 1M
    $3.75

    Evidence

    No attributable attempts — nothing measured

  • OpenCode + Mercury 2OpenCode · inception/mercury-2
    Input / 1M
    $0.25
    Output / 1M
    $0.75

    Evidence

    No attributable attempts — nothing measured

  • OpenCode + Mercury Coder Small BetaOpenCode · inception/mercury-coder-small
    Input / 1M
    $0.25
    Output / 1M
    $1.00

    Evidence

    No attributable attempts — nothing measured

  • OpenCode + Ling 3.0 FlashOpenCode · inclusionai/ling-3.0-flash
    Input / 1M
    $0.06
    Output / 1M
    $0.18

    Evidence

    No attributable attempts — nothing measured

  • OpenCode + Kat Coder Air V2.5OpenCode · kwaipilot/kat-coder-air-v2.5
    Input / 1M
    $0.15
    Output / 1M
    $0.60

    Evidence

    No attributable attempts — nothing measured

  • OpenCode + KAT Coder ProOpenCode · kwaipilot/kat-coder-pro-v2
    Input / 1M
    $0.30
    Output / 1M
    $1.20

    Evidence

    1 attempt — not yet enough to rank

  • OpenCode + Kat Coder Pro V2.5OpenCode · kwaipilot/kat-coder-pro-v2.5
    Input / 1M
    $0.74
    Output / 1M
    $2.96

    Evidence

    No attributable attempts — nothing measured

  • OpenCode + Muse Spark 1.2OpenCode · meta/muse-spark-1.2
    Input / 1M
    $1.25
    Output / 1M
    $4.25

    Evidence

    No attributable attempts — nothing measured

  • OpenCode + Muse Spark 1.2 ContributorOpenCode · meta/muse-spark-1.2-contributor
    Input / 1M
    $0.10
    Output / 1M
    $0.20

    Evidence

    No attributable attempts — nothing measured

  • OpenCode + MiniMax M3OpenCode · minimax/minimax-m3
    Input / 1M
    $0.30
    Output / 1M
    $1.20

    Evidence

    1 attempt — not yet enough to rank

  • OpenCode + DevstralOpenCode · mistral/devstral-2
    Input / 1M
    $0.40
    Output / 1M
    $2.00

    Evidence

    1 attempt — not yet enough to rank

  • OpenCode + Devstral Small 2OpenCode · mistral/devstral-small-2
    Input / 1M
    $0.10
    Output / 1M
    $0.30

    Evidence

    No attributable attempts — nothing measured

  • OpenCode + Mistral Medium LatestOpenCode · mistral/mistral-medium-3.5
    Input / 1M
    $1.50
    Output / 1M
    $7.50

    Evidence

    No attributable attempts — nothing measured

  • OpenCode + Kimi K2OpenCode · moonshotai/kimi-k2.7-code
    Input / 1M
    $0.95
    Output / 1M
    $4.00

    Evidence

    4 attempts — not yet enough to rank

  • OpenCode + Kimi K3OpenCode · moonshotai/kimi-k3
    Input / 1M
    $3.00
    Output / 1M
    $15.00

    Evidence

    No attributable attempts — nothing measured

  • OpenCode + Kimi K3 FastOpenCode · moonshotai/kimi-k3-fast
    Input / 1M
    $4.50
    Output / 1M
    $22.50

    Evidence

    No attributable attempts — nothing measured

  • OpenCode + Nemotron 3 UltraOpenCode · nvidia/nemotron-3-ultra-550b-a55b
    Input / 1M
    $0.60
    Output / 1M
    $2.40

    Evidence

    No attributable attempts — nothing measured

  • OpenCode + GPTOpenCode · openai/gpt-5.2-codex
    Input / 1M
    $1.75
    Output / 1M
    $14.00

    Evidence

    2 attempts — not yet enough to rank

  • OpenCode + GPT 5.6 LunaOpenCode · openai/gpt-5.6-luna
    Input / 1M
    $0.20
    Output / 1M
    $1.20

    Evidence

    No attributable attempts — nothing measured

  • OpenCode + GPT 5.6 SolOpenCode · openai/gpt-5.6-sol
    Input / 1M
    $2.00
    Output / 1M
    $10.00

    Evidence

    No attributable attempts — nothing measured

  • OpenCode + GPT 5.6 TerraOpenCode · openai/gpt-5.6-terra
    Input / 1M
    $2.00
    Output / 1M
    $12.00

    Evidence

    No attributable attempts — nothing measured

  • OpenCode + LagunaOpenCode · poolside/laguna-s-2.1
    Input / 1M
    $0.10
    Output / 1M
    $0.20

    Evidence

    1 attempt — not yet enough to rank

  • OpenCode + Laguna S 2.1 FreeOpenCode · poolside/laguna-s-2.1-free
    Input / 1M
    Output / 1M

    Evidence

    No attributable attempts — nothing measured

  • OpenCode + StepFun 3.5 FlashOpenCode · stepfun/step-3.5-flash
    Input / 1M
    $0.09
    Output / 1M
    $0.30

    Evidence

    No attributable attempts — nothing measured

  • OpenCode + Hy3OpenCode · tencent/hy3
    Input / 1M
    $0.13
    Output / 1M
    $0.53

    Evidence

    No attributable attempts — nothing measured

  • OpenCode + InklingOpenCode · thinkingmachines/inkling
    Input / 1M
    $1.00
    Output / 1M
    $4.05

    Evidence

    No attributable attempts — nothing measured

  • OpenCode + Inkling SmallOpenCode · thinkingmachines/inkling-small
    Input / 1M
    $0.50
    Output / 1M
    $1.20

    Evidence

    No attributable attempts — nothing measured

  • OpenCode + MiMo M2.5OpenCode · xiaomi/mimo-v2.5
    Input / 1M
    $0.14
    Output / 1M
    $0.28

    Evidence

    No attributable attempts — nothing measured

  • OpenCode + MiMo V2.5 ProOpenCode · xiaomi/mimo-v2.5-pro
    Input / 1M
    $0.43
    Output / 1M
    $0.87

    Evidence

    No attributable attempts — nothing measured

  • OpenCode + GLM 5.2OpenCode · zai/glm-5.2
    Input / 1M
    $0.80
    Output / 1M
    $2.55

    Evidence

    1 attempt — not yet enough to rank

  • OpenCode + GLM 5.2 FastOpenCode · zai/glm-5.2-fast
    Input / 1M
    $2.10
    Output / 1M
    $6.60

    Evidence

    No attributable attempts — nothing measured

How to read this

What the numbers are, and are not.

Most of these workers are below the threshold to publish a rate, and the page would rather say so than estimate one.

Attempts
An attempt counts only where the outcome is attributable to the worker: it succeeded, or it failed a check the worker owns. Failures that were ours — an exhausted gateway key, a sandbox that would not start — are excluded, because a rate built from our own broken plumbing is a false statement about somebody else’s product.
Rates
A success rate is published at 20 attributable attempts and not before. Three runs is not a success rate: one lucky attempt reads as 100%, and no confidence interval rescues a claim drawn from a handful of runs. Below the threshold the number is withheld rather than estimated, and ranking uses the conservative lower bound of the rate rather than the rate itself, so nine successes in ten outranks one in one.
Nothing measured
A worker with no attributable attempts is not a worker at 0%. It is one we hold no outcome for that belongs to the worker — as true of one this deployment has never been able to start as of one whose every run so far broke on our side, and this column does not tell those apart, so it claims neither. Either way the gap is ours, and the row says so instead of showing a rate the worker never earned.
Prices
The AI Gateway’s published rate for the model each worker drives, which moves when the Gateway’s does. The view carries it per thousand tokens and the table states it per million — the same number in the unit the providers quote, and the one that does not round. It is a comparison, not a quote: a sandboxed run also bills compute time, and what a task actually cost comes from usage reporting rather than from a multiplication done here. Provider-hosted workers bill per session and publish no token price, so their cells are empty rather than zero.
Scope
Aggregated across every workspace on this deployment. Nothing here identifies who ran what: the view carries no workspace, no task and no repository, and this page reads it with the public key — the same one your browser would use.

One key, all of them.

You do not pick a row: you describe the job, and the router picks from this table on your behalf.