Muse Spark 1.3 Contributor run record

Muse Spark 1.3 Contributor

meta/muse-spark-1.3-contributor

starter@v1HostedOpenRouterserved by Meta1M context2026-10-04 21:09 UTC

Overall score

108 of 153 test cases fully passed

Throughput

156.9tok/s

Mean per test case, 107 tok/s over the run

Output per test

8.7Ktokens

Answer plus reasoning

Reasoning per test

7.9Ktokens

91% of the output

Total tokens

2.8M

1.5M in, 1.3M out

Run time

4h 19m

1m 42s per caseRequest time, summed

Cost

$0.41

$0.0027 per caseAt the provider's list price

Every test case

153 test cases in 8 tasks

One square per test case. Select a square to read the answer.

PassedPartial creditFailed

57Agents12 of 21 passed

12 of 21 passed

90Structured Extraction19 of 21 passed

19 of 21 passed

48Coding10 of 21 passed

10 of 21 passed

81Grounded Answers17 of 21 passed

17 of 21 passed

84Instruction Following17 of 21 passed

17 of 21 passed

76Business Logic16 of 21 passed

16 of 21 passed

76Security & permissions16 of 21 passed

16 of 21 passed

17Diagnosis & investigation1 of 6 passed

1 of 6 passed

Speed and output per test case

One bar per test case. Hover to inspect a case, select it to read the answer.

ThroughputPeak 557 tok/s

Output tokensReasoningAnswer

Test by test

Every answer with its reasoning and the checks it had to pass.

Reproduce this run

The prompts stay private so models can't train on them. The hashes pin the exact suite and config this run used.

Serve the modelOpenRouterHosted

curl https://openrouter.ai/api/v1/chat/completions \
  -H "Authorization: Bearer $OPENROUTER_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model":"meta/muse-spark-1.3-contributor","provider":{"order":["meta"],"allow_fallbacks":false},"service_tier":"default","messages":[{"role":"user","content":"Hello"}]}'

The request the benchmark sent, pinned to the provider that served it.

Record id
Suite hash, starter@v1
Config hash

Configuration

Where the suite ran, and the token budget each task got.

Deployment: Hosted
API: OpenRouter
Serving provider: Meta
Precision: Provider default
Context window: 1M
Temperature: 0
Bench version: v1.0.0
Ran at: 2026-10-04 21:09 UTC

Provider parameters

{
  "provider": {
    "order": [
      "meta"
    ],
    "allow_fallbacks": false
  },
  "service_tier": "default"
}

Token budget per task
Task	Reasoning	Max output per call
Agents	4.1Khigh	12.3K
Structured Extraction	4.1Khigh	12.3K
Coding	4.1Khigh	12.3K
Grounded Answers	4.1Khigh	12.3K
Instruction Following	4.1Khigh	12.3K
Business Logic	4.1Khigh	12.3K
Security & permissions	4.1Khigh	12.3K
Diagnosis & investigation	4.1Khigh	12.3K

Muse Spark 1.3 Contributor Compare with another model

Every test case

Speed and output per test case

Test by test

57AgentsRuns multi-step jobs through tools, within a written policy12 of 21 passedp90 latency 8m 42s

90Structured ExtractionTurns logs, diffs and records into exact JSON19 of 21 passedp90 latency 1m 21s

48CodingWrites code and SQL that pass hidden tests10 of 21 passedp90 latency 3m 5s

81Grounded AnswersAnswers from the given documents with sources, never guesses17 of 21 passedp90 latency 1m 12s

84Instruction FollowingApplies every rule, including ones changed mid-conversation17 of 21 passedp90 latency 2m 5s

76Business LogicComputes billing, scheduling and pricing rules exactly16 of 21 passedp90 latency 1m 47s

76Security & permissionsIgnores planted instructions, still does the allowed work16 of 21 passedp90 latency 5m 57s

17Diagnosis & investigationChooses diagnostic checks, repairs supported causes and verifies recovery1 of 6 passedp90 latency 5m 52s

Reproduce this run

Configuration