Mistral Large 4 vs Muse Spark Open
Wins 3 of 7 areas
Coding · Agents · Long documents
Wins 4 of 7 areas
Reasoning · Facts · Images and charts · Math
Muse Spark Open wins more areas, narrowly.Mistral Large 4 is cheaper and better at agents.
Scores updated · 27 tests both models report · How we compare
Where each one wins
Tests won in each of the seven areas where both have results. Each piece is one test, so a longer bar means more evidence; grey means the two scored within a point of each other.
- FactsGetting facts right instead of making them up03Muse Spark Open3 of 3 tests
- ReasoningHard problems that need careful thinking02Muse Spark Open2 of 2 tests
- MathCompetition and research-level math02Muse Spark Open2 of 2 tests
- Images and chartsUnderstanding pictures, charts and video01Muse Spark Open1 of 1 test
- AgentsCarrying out multi-step tasks on its own20Mistral Large 42 of 2 tests
- CodingWriting and fixing software21Mistral Large 42 of 3 tests
- Long documentsFinding answers in very long texts10Mistral Large 41 of 1 test
Images and charts and long documents rest on a single test each.
What it costs
Prices per million tokens, roughly 750,000 words. The bars show the cost of a million tokens read plus a million written.
Mistral Large 4 costs 49% less for the same work.
The biggest differences
The three tests each model wins by the widest margin. Scores are out of 100.
Where Muse Spark Open pulls ahead
- Short factual questions, answered correctlySimpleQA Verified+34.2points ahead
- Avoids making up answers it doesn't knowAA-Omniscience · Non-hallucination+17.8points ahead
- Answers hard knowledge questions correctlyAA-Omniscience · Accuracy+13.5points ahead
Where Mistral Large 4 pulls ahead
- Complex command-line tasks across many fieldsTerminal-Bench 4.0+16.6points ahead
- Command-line tasks in a real terminalTerminal-Bench 2.1+13.8points ahead
- Reasons across sets of long documentsAA-LCR+3points ahead
Every test, side by side
All 27 tests both models report. The winning score is in its model's colour; marks a score checked independently.
CodingMistral Large 4
- Terminal-Bench 2.1Mistral Large 4 by 13.890.276.4+13.8
- LMArena · WebDevMistral Large 4 by 43 rating points15701527+43 rating
- SciCodeMuse Spark Open by 4.254.258.4+4.2
AgentsMistral Large 4
- Terminal-Bench 4.0Mistral Large 4 by 16.624.78.1+16.6
- GDPValMistral Large 4 by 1.146.245.1+1.1
ReasoningMuse Spark Open
- Humanity's Last ExamMuse Spark Open by 8.93543.9+8.9
- CritPtMuse Spark Open by 4.91014.9+4.9
FactsMuse Spark Open
- SimpleQA VerifiedMuse Spark Open by 34.222.456.6+34.2
- AA-Omniscience · Non-hallucinationMuse Spark Open by 17.858.175.9+17.8
- AA-Omniscience · AccuracyMuse Spark Open by 13.525.839.3+13.5
Images and chartsMuse Spark Open
- MMMU-ProMuse Spark Open by 4.576.480.9+4.5
MathMuse Spark Open
- FrontierMath Tiers 1-3 (v2)Muse Spark Open by 11.258.669.8+11.2
- FrontierMath Tier 4Muse Spark Open by 4.926.831.7+4.9
Other results13 tests, not counted
Tests outside the eight areas. They are not counted above: several are summary scores built from other tests, or the same test under another name.
- Terminal-Bench 4.0 (AA Coding Agent Index)Mistral Large 4 by 35416+35
- AA-OmniscienceMuse Spark Open by 30-5.324.6+30
- DeepSWE 1.1Muse Spark Open by 15.361.777+15.3
- HLE-DiamondMuse Spark Open by 917.326.3+9
- BeyondAIMEMistral Large 4 by 5.790.684.9+5.7
- JailbreakBench (cyber subset): safe-response rateMuse Spark Open by 2.797.3100+2.7
- CritPt v2 HardMuse Spark Open by 2.52729.5+2.5
- Terminal-Bench-Science 0.1Mistral Large 4 by 2.44.31.9+2.4
- StrongREJECT (cyber subset): refusal rateMistral Large 4 by 1.992.690.7+1.9
- CritPt v2 (Hard Subset)Muse Spark Open by 1.52627.5+1.5
- GDP (Surge AI)Muse Spark Open by 117.618.6+1
- AgentHarm (cyber subset): 100 minus harm scoretie10099.4tie
- AA Intelligencetie38.438.9tie
Questions people ask
Which is better, Mistral Large 4 or Muse Spark Open?
Muse Spark Open wins four of the seven areas where both have results: reasoning, facts, images and charts and math. Mistral Large 4 wins coding, agents and long documents, and costs 49% less.
Which is better for coding?
Mistral Large 4. It wins 2 of the 3 coding tests both models report; Muse Spark Open wins 1.
Which is cheaper?
Mistral Large 4 costs $0.70 per million input tokens and $2.10 per million output tokens; Muse Spark Open costs $1.25 and $4.25. That makes Mistral Large 4 about 49% cheaper for the same work.
How do you compare the two?
We use the 27 benchmark tests both models have published scores on. The verdict counts the 14 tests in the eight capability areas, and a gap under one point (ten on rating-style scales) counts as a tie. The other 13 are listed but not counted, because several are summary scores or repeat a test. Each score is the one shown on the model's own page.